4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
27 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
29 The POLY1305 MAC (message authentication code) implementation in OpenSSL
30 does not save the contents of non-volatile XMM registers on Windows 64
31 platform when calculating the MAC of data larger than 64 bytes. Before
32 returning to the caller all the XMM registers are set to zero rather than
33 restoring their previous content. The vulnerable code is used only on newer
34 x86_64 processors supporting the AVX512-IFMA instructions.
36 The consequences of this kind of internal application state corruption can
37 be various - from no consequences, if the calling application does not
38 depend on the contents of non-volatile XMM registers at all, to the worst
39 consequences, where the attacker could get complete control of the
40 application process. However given the contents of the registers are just
41 zeroized so the attacker cannot put arbitrary values inside, the most likely
42 consequence, if any, would be an incorrect result of some application
43 dependent calculations or a crash leading to a denial of service.
49 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
51 * Fix excessive time spent checking DH q parameter value.
53 The function DH_check() performs various checks on DH parameters. After
54 fixing CVE-2023-3446 it was discovered that a large q parameter value can
55 also trigger an overly long computation during some of these checks.
56 A correct q value, if present, cannot be larger than the modulus p
57 parameter, thus it is unnecessary to perform these checks if q is larger
60 If DH_check() is called with such q parameter value,
61 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
62 intensive checks are skipped.
68 * Fix DH_check() excessive time with over sized modulus.
70 The function DH_check() performs various checks on DH parameters. One of
71 those checks confirms that the modulus ("p" parameter) is not too large.
72 Trying to use a very large modulus is slow and OpenSSL will not normally use
73 a modulus which is over 10,000 bits in length.
75 However the DH_check() function checks numerous aspects of the key or
76 parameters that have been supplied. Some of those checks use the supplied
77 modulus value even if it has already been found to be too large.
79 A new limit has been added to DH_check of 32,768 bits. Supplying a
80 key/parameters with a modulus over this size will simply cause DH_check() to
87 * Do not ignore empty associated data entries with AES-SIV.
89 The AES-SIV algorithm allows for authentication of multiple associated
90 data entries along with the encryption. To authenticate empty data the
91 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
92 with NULL pointer as the output buffer and 0 as the input buffer length.
93 The AES-SIV implementation in OpenSSL just returns success for such call
94 instead of performing the associated data authentication operation.
95 The empty data thus will not be authenticated. ([CVE-2023-2975])
97 Thanks to Juerg Wullschleger (Google) for discovering the issue.
99 The fix changes the authentication tag value and the ciphertext for
100 applications that use empty associated data entries with AES-SIV.
101 To decrypt data encrypted with previous versions of OpenSSL the application
102 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
107 * When building with the `enable-fips` option and using the resulting
108 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
109 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
110 not operate with truncated digests (FIPS 140-3 IG G.R).
114 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
116 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
117 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
119 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
120 numeric text form. For gigantic sub-identifiers, this would take a very
121 long time, the time complexity being O(n^2) where n is the size of that
122 sub-identifier. ([CVE-2023-2650])
124 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
125 IDENTIFIER to canonical numeric text form if the size of that OBJECT
126 IDENTIFIER is 586 bytes or less, and fail otherwise.
128 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
129 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
130 most 128 sub-identifiers, and that the maximum value that each sub-
131 identifier may have is 2^32-1 (4294967295 decimal).
133 For each byte of every sub-identifier, only the 7 lower bits are part of
134 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
135 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
140 * Multiple algorithm implementation fixes for ARM BE platforms.
144 * Added a -pedantic option to fipsinstall that adjusts the various
145 settings to ensure strict FIPS compliance rather than backwards
150 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
151 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
152 trigger a crash of an application using AES-XTS decryption if the memory
153 just after the buffer being decrypted is not mapped.
154 Thanks to Anton Romanov (Amazon) for discovering the issue.
159 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
160 The previous fix for this timing side channel turned out to cause
161 a severe 2-3x performance regression in the typical use case
162 compared to 3.0.7. The new fix uses existing constant time
163 code paths, and restores the previous performance level while
164 fully eliminating all existing timing side channels.
165 The fix was developed by Bernd Edlinger with testing support
170 * Add FIPS provider configuration option to disallow the use of
171 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
172 The option '-no_drbg_truncated_digests' can optionally be
173 supplied to 'openssl fipsinstall'.
177 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
178 that it does not enable policy checking. Thanks to David Benjamin for
179 discovering this issue.
184 * Fixed an issue where invalid certificate policies in leaf certificates are
185 silently ignored by OpenSSL and other certificate policy checks are skipped
186 for that certificate. A malicious CA could use this to deliberately assert
187 invalid certificate policies in order to circumvent policy checking on the
188 certificate altogether.
193 * Limited the number of nodes created in a policy tree to mitigate
194 against CVE-2023-0464. The default limit is set to 1000 nodes, which
195 should be sufficient for most installations. If required, the limit
196 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
197 time define to a desired maximum number of nodes or zero to allow
203 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
205 * Add FIPS provider configuration option to enforce the
206 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
207 The option '-ems-check' can optionally be supplied to
208 'openssl fipsinstall'.
212 * The FIPS provider includes a few non-approved algorithms for
213 backward compatibility purposes and the "fips=yes" property query
214 must be used for all algorithm fetches to ensure FIPS compliance.
216 The algorithms that are included but not approved are Triple DES ECB,
217 Triple DES CBC and EdDSA.
221 * Added support for KMAC in KBKDF.
225 * RNDR and RNDRRS support in provider functions to provide
226 random number generation for Arm CPUs (aarch64).
230 * s_client and s_server apps now explicitly say when the TLS version
231 does not include the renegotiation mechanism. This avoids confusion
232 between that scenario versus when the TLS version includes secure
233 renegotiation but the peer lacks support for it.
237 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
239 *Tomasz Kantecki, Andrey Matyukov*
241 * The various OBJ_* functions have been made thread safe.
245 * Parallel dual-prime 1536/2048-bit modular exponentiation for
246 AVX512_IFMA capable processors.
248 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
250 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
251 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
252 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
253 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
254 `OPENSSL_NO_DEPRECATED_3_1`.
256 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
257 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
258 definitions for these functions regardless of whether
259 `OPENSSL_NO_DEPRECATED_3_1` is defined.
261 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
262 functions regardless of whether they are using them. It is recommended that
263 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
267 * When generating safe-prime DH parameters set the recommended private key
268 length equivalent to minimum key lengths as in RFC 7919.
272 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
273 maximum size that is smaller or equal to the digest length to comply with
274 FIPS 186-4 section 5. This is implemented by a new option
275 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
276 `rsa_pss_saltlen` parameter, which is now the default. Signature
277 verification is not affected by this change and continues to work as before.
284 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
285 listed here are only a brief description.
286 The migration guide contains more detailed information related to new features,
287 breaking changes, and mappings for the large list of deprecated functions.
289 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
291 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
293 * Fixed NULL dereference during PKCS7 data verification.
295 A NULL pointer can be dereferenced when signatures are being
296 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
297 algorithm used for the signature is known to the OpenSSL library but
298 the implementation of the hash algorithm is not available the digest
299 initialization will fail. There is a missing check for the return
300 value from the initialization function which later leads to invalid
301 usage of the digest API most likely leading to a crash.
304 PKCS7 data is processed by the SMIME library calls and also by the
305 time stamp (TS) library calls. The TLS implementation in OpenSSL does
306 not call these functions however third party applications would be
307 affected if they call these functions to verify signatures on untrusted
312 * Fixed X.400 address type confusion in X.509 GeneralName.
314 There is a type confusion vulnerability relating to X.400 address processing
315 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
316 but the public structure definition for GENERAL_NAME incorrectly specified
317 the type of the x400Address field as ASN1_TYPE. This field is subsequently
318 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
321 When CRL checking is enabled (i.e. the application sets the
322 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
323 pass arbitrary pointers to a memcmp call, enabling them to read memory
324 contents or enact a denial of service.
329 * Fixed NULL dereference validating DSA public key.
331 An invalid pointer dereference on read can be triggered when an
332 application tries to check a malformed DSA public key by the
333 EVP_PKEY_public_check() function. This will most likely lead
334 to an application crash. This function can be called on public
335 keys supplied from untrusted sources which could allow an attacker
336 to cause a denial of service attack.
338 The TLS implementation in OpenSSL does not call this function
339 but applications might call the function if there are additional
340 security requirements imposed by standards such as FIPS 140-3.
343 *Shane Lontis, Tomáš Mráz*
345 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
347 An invalid pointer dereference on read can be triggered when an
348 application tries to load malformed PKCS7 data with the
349 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
351 The result of the dereference is an application crash which could
352 lead to a denial of service attack. The TLS implementation in OpenSSL
353 does not call this function however third party applications might
354 call these functions on untrusted data.
359 * Fixed Use-after-free following BIO_new_NDEF.
361 The public API function BIO_new_NDEF is a helper function used for
362 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
363 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
364 be called directly by end user applications.
366 The function receives a BIO from the caller, prepends a new BIO_f_asn1
367 filter BIO onto the front of it to form a BIO chain, and then returns
368 the new head of the BIO chain to the caller. Under certain conditions,
369 for example if a CMS recipient public key is invalid, the new filter BIO
370 is freed and the function returns a NULL result indicating a failure.
371 However, in this case, the BIO chain is not properly cleaned up and the
372 BIO passed by the caller still retains internal pointers to the previously
373 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
374 then a use-after-free will occur. This will most likely result in a crash.
377 *Viktor Dukhovni, Matt Caswell*
379 * Fixed Double free after calling PEM_read_bio_ex.
381 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
382 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
383 data. If the function succeeds then the "name_out", "header" and "data"
384 arguments are populated with pointers to buffers containing the relevant
385 decoded data. The caller is responsible for freeing those buffers. It is
386 possible to construct a PEM file that results in 0 bytes of payload data.
387 In this case PEM_read_bio_ex() will return a failure code but will populate
388 the header argument with a pointer to a buffer that has already been freed.
389 If the caller also frees this buffer then a double free will occur. This
390 will most likely lead to a crash.
392 The functions PEM_read_bio() and PEM_read() are simple wrappers around
393 PEM_read_bio_ex() and therefore these functions are also directly affected.
395 These functions are also called indirectly by a number of other OpenSSL
396 functions including PEM_X509_INFO_read_bio_ex() and
397 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
398 internal uses of these functions are not vulnerable because the caller does
399 not free the header argument if PEM_read_bio_ex() returns a failure code.
402 *Kurt Roeckx, Matt Caswell*
404 * Fixed Timing Oracle in RSA Decryption.
406 A timing based side channel exists in the OpenSSL RSA Decryption
407 implementation which could be sufficient to recover a plaintext across
408 a network in a Bleichenbacher style attack. To achieve a successful
409 decryption an attacker would have to be able to send a very large number
410 of trial messages for decryption. The vulnerability affects all RSA padding
411 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
414 *Dmitry Belyavsky, Hubert Kario*
416 * Fixed X.509 Name Constraints Read Buffer Overflow.
418 A read buffer overrun can be triggered in X.509 certificate verification,
419 specifically in name constraint checking. The read buffer overrun might
420 result in a crash which could lead to a denial of service attack.
421 In a TLS client, this can be triggered by connecting to a malicious
422 server. In a TLS server, this can be triggered if the server requests
423 client authentication and a malicious client connects.
428 * Fixed X.509 Policy Constraints Double Locking security issue.
430 If an X.509 certificate contains a malformed policy constraint and
431 policy processing is enabled, then a write lock will be taken twice
432 recursively. On some operating systems (most widely: Windows) this
433 results in a denial of service when the affected process hangs. Policy
434 processing being enabled on a publicly facing server is not considered
435 to be a common setup.
440 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
441 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
442 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
443 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
444 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
445 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
446 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
447 for legacy EC and SM2 keys is also changed similarly to honor the
448 equivalent conversion format flag as specified in the underlying
449 `EC_KEY` object being exported to a provider, when this function is
450 called through `EVP_PKEY_export()`.
454 * Fixed a type confusion vulnerability relating to X.400 address processing
455 inside an X.509 GeneralName. X.400 addresses were parsed as an `ASN1_STRING`
456 but subsequently interpreted by `GENERAL_NAME_cmp` as an `ASN1_TYPE`. This
457 vulnerability may allow an attacker who can provide a certificate chain and
458 CRL (neither of which need have a valid signature) to pass arbitrary pointers
459 to a `memcmp` call, creating a possible read primitive, subject to some
460 constraints. Refer to the advisory for more information. Thanks to David
461 Benjamin for discovering this issue. ([CVE-2023-0286])
463 This issue has been fixed by changing the public header file definition of
464 `GENERAL_NAME` so that `x400Address` reflects the implementation. It was not
465 possible for any existing application to successfully use the existing
466 definition; however, if any application references the `x400Address` field
467 (e.g. in dead code), note that the type of this field has changed. There is
472 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
474 * Fixed two buffer overflows in punycode decoding functions.
476 A buffer overrun can be triggered in X.509 certificate verification,
477 specifically in name constraint checking. Note that this occurs after
478 certificate chain signature verification and requires either a CA to
479 have signed the malicious certificate or for the application to continue
480 certificate verification despite failure to construct a path to a trusted
483 In a TLS client, this can be triggered by connecting to a malicious
484 server. In a TLS server, this can be triggered if the server requests
485 client authentication and a malicious client connects.
487 An attacker can craft a malicious email address to overflow
488 an arbitrary number of bytes containing the `.` character (decimal 46)
489 on the stack. This buffer overflow could result in a crash (causing a
493 An attacker can craft a malicious email address to overflow four
494 attacker-controlled bytes on the stack. This buffer overflow could
495 result in a crash (causing a denial of service) or potentially remote code
496 execution depending on stack layout for any given platform/compiler.
501 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
502 parameters in OpenSSL code.
503 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
504 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
505 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
506 Using these invalid names may cause algorithms to use slower methods
507 that ignore the CRT parameters.
511 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
516 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
517 data to be signed before signing the certificate.
521 * Added RIPEMD160 to the default provider.
525 * Ensured that the key share group sent or accepted for the key exchange
526 is allowed for the protocol version.
530 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
532 * OpenSSL supports creating a custom cipher via the legacy
533 EVP_CIPHER_meth_new() function and associated function calls. This function
534 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
535 to use the new provider mechanism in order to implement custom ciphers.
537 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
538 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
539 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
540 and decryption initialisation functions). Instead of using the custom cipher
541 directly it incorrectly tries to fetch an equivalent cipher from the
542 available providers. An equivalent cipher is found based on the NID passed to
543 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
544 given cipher. However it is possible for an application to incorrectly pass
545 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
546 is used in this way the OpenSSL encryption/decryption initialisation function
547 will match the NULL cipher as being equivalent and will fetch this from the
548 available providers. This will succeed if the default provider has been
549 loaded (or if a third party provider has been loaded that offers this
550 cipher). Using the NULL cipher means that the plaintext is emitted as the
553 Applications are only affected by this issue if they call
554 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
555 encryption/decryption initialisation function. Applications that only use
556 SSL/TLS are not impacted by this issue.
561 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
566 * Fixed the linux-mips64 Configure target which was missing the
567 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
572 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
577 * Correctly handle a retransmitted ClientHello in DTLS
581 * Fixed detection of ktls support in cross-compile environment on Linux
585 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
590 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
591 report correct results in some cases
595 * Fix UWP builds by defining VirtualLock
599 * For known safe primes use the minimum key length according to RFC 7919.
600 Longer private key sizes unnecessarily raise the cycles needed to compute the
601 shared secret without any increase of the real security. This fixes a
602 regression from 1.1.1 where these shorter keys were generated for the known
607 * Added the loongarch64 target
611 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
612 only passed to the FIPS provider and not to the default or legacy provider.
616 * Fixed reported performance degradation on aarch64. Restored the
617 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
618 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
619 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
620 The new algorithm is still used for 32 bit targets.
624 * Added a missing header for memcmp that caused compilation failure on some
629 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
631 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
632 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
633 This issue makes the RSA implementation with 2048 bit private keys
634 incorrect on such machines and memory corruption will happen during
635 the computation. As a consequence of the memory corruption an attacker
636 may be able to trigger a remote code execution on the machine performing
639 SSL/TLS servers or other servers using 2048 bit RSA private keys running
640 on machines supporting AVX512IFMA instructions of the X86_64 architecture
641 are affected by this issue.
646 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
647 implementation would not encrypt the entirety of the data under some
648 circumstances. This could reveal sixteen bytes of data that was
649 preexisting in the memory that wasn't written. In the special case of
650 "in place" encryption, sixteen bytes of the plaintext would be revealed.
652 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
653 they are both unaffected.
656 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
658 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
660 * In addition to the c_rehash shell command injection identified in
661 CVE-2022-1292, further bugs where the c_rehash script does not
662 properly sanitise shell metacharacters to prevent command injection have been
665 When the CVE-2022-1292 was fixed it was not discovered that there
666 are other places in the script where the file names of certificates
667 being hashed were possibly passed to a command executed through the shell.
669 This script is distributed by some operating systems in a manner where
670 it is automatically executed. On such operating systems, an attacker
671 could execute arbitrary commands with the privileges of the script.
673 Use of the c_rehash script is considered obsolete and should be replaced
674 by the OpenSSL rehash command line tool.
677 *Daniel Fiala, Tomáš Mráz*
679 * Case insensitive string comparison no longer uses locales. It has instead
680 been directly implemented.
684 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
686 * Case insensitive string comparison is reimplemented via new locale-agnostic
687 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
688 comparison. The previous implementation had problems when the Turkish locale
693 * Fixed a bug in the c_rehash script which was not properly sanitising shell
694 metacharacters to prevent command injection. This script is distributed by
695 some operating systems in a manner where it is automatically executed. On
696 such operating systems, an attacker could execute arbitrary commands with the
697 privileges of the script.
699 Use of the c_rehash script is considered obsolete and should be replaced
700 by the OpenSSL rehash command line tool.
705 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
706 certificate on an OCSP response. The bug caused the function in the case
707 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
708 response (meaning a successful verification) even in the case where the
709 response signing certificate fails to verify.
711 It is anticipated that most users of `OCSP_basic_verify` will not use the
712 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
713 a negative value (indicating a fatal error) in the case of a certificate
714 verification failure. The normal expected return value in this case would be
717 This issue also impacts the command line OpenSSL "ocsp" application. When
718 verifying an ocsp response with the "-no_cert_checks" option the command line
719 application will report that the verification is successful even though it
720 has in fact failed. In this case the incorrect successful response will also
721 be accompanied by error messages showing the failure and contradicting the
722 apparently successful result.
727 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
728 AAD data as the MAC key. This made the MAC key trivially predictable.
730 An attacker could exploit this issue by performing a man-in-the-middle attack
731 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
732 that the modified data would still pass the MAC integrity check.
734 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
735 endpoint will always be rejected by the recipient and the connection will
736 fail at that point. Many application protocols require data to be sent from
737 the client to the server first. Therefore, in such a case, only an OpenSSL
738 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
740 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
741 sent in both directions. In this case both clients and servers could be
742 affected, regardless of the application protocol.
744 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
745 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
746 the handshake when using this ciphersuite.
748 The confidentiality of data is not impacted by this issue, i.e. an attacker
749 cannot decrypt data that has been encrypted using this ciphersuite - they can
752 In order for this attack to work both endpoints must legitimately negotiate
753 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
754 OpenSSL 3.0, and is not available within the default provider or the default
755 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
756 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
757 following must have occurred:
759 1) OpenSSL must have been compiled with the (non-default) compile time option
760 enable-weak-ssl-ciphers
762 2) OpenSSL must have had the legacy provider explicitly loaded (either
763 through application code or via configuration)
765 3) The ciphersuite must have been explicitly added to the ciphersuite list
767 4) The libssl security level must have been set to 0 (default is 1)
769 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
771 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
772 others that both endpoints have in common
777 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
778 occupied by the removed hash table entries.
780 This function is used when decoding certificates or keys. If a long lived
781 process periodically decodes certificates or keys its memory usage will
782 expand without bounds and the process might be terminated by the operating
783 system causing a denial of service. Also traversing the empty hash table
784 entries will take increasingly more time.
786 Typically such long lived processes might be TLS clients or TLS servers
787 configured to accept client certificate authentication.
790 *Hugo Landau, Aliaksei Levin*
792 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
793 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
794 statistics are no longer supported. For compatibility, these statistics are
795 still listed in the output but are now always reported as zero.
799 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
801 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
802 for non-prime moduli.
804 Internally this function is used when parsing certificates that contain
805 elliptic curve public keys in compressed form or explicit elliptic curve
806 parameters with a base point encoded in compressed form.
808 It is possible to trigger the infinite loop by crafting a certificate that
809 has invalid explicit curve parameters.
811 Since certificate parsing happens prior to verification of the certificate
812 signature, any process that parses an externally supplied certificate may thus
813 be subject to a denial of service attack. The infinite loop can also be
814 reached when parsing crafted private keys as they can contain explicit
815 elliptic curve parameters.
817 Thus vulnerable situations include:
819 - TLS clients consuming server certificates
820 - TLS servers consuming client certificates
821 - Hosting providers taking certificates or private keys from customers
822 - Certificate authorities parsing certification requests from subscribers
823 - Anything else which parses ASN.1 elliptic curve parameters
825 Also any other applications that use the BN_mod_sqrt() where the attacker
826 can control the parameter values are vulnerable to this DoS issue.
831 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
832 to the list of ciphersuites providing Perfect Forward Secrecy as
833 required by SECLEVEL >= 3.
835 *Dmitry Belyavskiy, Nicola Tuveri*
837 * Made the AES constant time code for no-asm configurations
838 optional due to the resulting 95% performance degradation.
839 The AES constant time code can be enabled, for no assembly
840 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
844 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
849 * The negative return value handling of the certificate verification callback
850 was reverted. The replacement is to set the verification retry state with
851 the SSL_set_retry_verify() function.
855 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
857 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
858 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
859 verify a certificate supplied by a server. That function may return a
860 negative return value to indicate an internal error (for example out of
861 memory). Such a negative return value is mishandled by OpenSSL and will cause
862 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
863 success and a subsequent call to SSL_get_error() to return the value
864 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
865 returned by OpenSSL if the application has previously called
866 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
867 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
868 totally unexpected and applications may not behave correctly as a result. The
869 exact behaviour will depend on the application but it could result in
870 crashes, infinite loops or other similar incorrect responses.
872 This issue is made more serious in combination with a separate bug in OpenSSL
873 3.0 that will cause X509_verify_cert() to indicate an internal error when
874 processing a certificate chain. This will occur where a certificate does not
875 include the Subject Alternative Name extension but where a Certificate
876 Authority has enforced name constraints. This issue can occur even with valid
882 * Corrected a few file name and file reference bugs in the build,
883 installation and setup scripts, which lead to installation verification
884 failures. Slightly enhanced the installation verification script.
888 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
893 * Fixed PVK encoder to properly query for the passphrase.
897 * Multiple fixes in the OSSL_HTTP API functions.
901 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
902 OSSL_PARAM_INTEGER data type and return error on negative numbers
903 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
904 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
908 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
912 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
916 * Multiple threading fixes.
920 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
924 * Allow fetching an operation from the provider that owns an unexportable key
925 as a fallback if that is still allowed by the property query.
929 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
931 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
936 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
937 S390X capability vector to zero. This simplifies testing of different code
938 paths on S390X architecture.
942 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
943 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
944 SP 800-38D". The communication will fail at this point.
948 * The EC_GROUP_clear_free() function is deprecated as there is nothing
949 confidential in EC_GROUP data.
953 * The byte order mark (BOM) character is ignored if encountered at the
954 beginning of a PEM-formatted file.
958 * Added CMS support for the Russian GOST algorithms.
962 * Due to move of the implementation of cryptographic operations
963 to the providers, validation of various operation parameters can
964 be postponed until the actual operation is executed where previously
965 it happened immediately when an operation parameter was set.
967 For example when setting an unsupported curve with
968 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
969 fail but later keygen operations with the EVP_PKEY_CTX will fail.
971 *OpenSSL team members and many third party contributors*
973 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
974 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
975 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
976 instead to retrieve these algorithms from a provider.
980 * On build targets where the multilib postfix is set in the build
981 configuration the libdir directory was changing based on whether
982 the lib directory with the multilib postfix exists on the system
983 or not. This unpredictable behavior was removed and eventual
984 multilib postfix is now always added to the default libdir. Use
985 `--libdir=lib` to override the libdir if adding the postfix is
990 * The triple DES key wrap functionality now conforms to RFC 3217 but is
991 no longer interoperable with OpenSSL 1.1.1.
995 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
996 function codes, this function can only cause problems for calling
1001 * Add a configurable flag to output date formats as ISO 8601. Does not
1002 change the default date format.
1006 * Version of MSVC earlier than 1300 could get link warnings, which could
1007 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1008 Support for this flag has been removed.
1012 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1013 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1014 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1015 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1016 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1020 * The signatures of the functions to get and set options on SSL and
1021 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1022 Some source code changes may be required.
1026 * The public definitions of conf_method_st and conf_st have been
1027 deprecated. They will be made opaque in a future release.
1029 *Rich Salz and Tomáš Mráz*
1031 * Client-initiated renegotiation is disabled by default. To allow it, use
1032 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1033 flag, or the "ClientRenegotiation" config parameter as appropriate.
1037 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1038 or modify relative pathname inclusion.
1042 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1043 validated. Please consult the README-FIPS and
1044 README-PROVIDERS files, as well as the migration guide.
1046 *OpenSSL team members and many third party contributors*
1048 * For the key types DH and DHX the allowed settable parameters are now different.
1052 * The openssl commands that read keys, certificates, and CRLs now
1053 automatically detect the PEM or DER format of the input files.
1055 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1057 * Added enhanced PKCS#12 APIs which accept a library context.
1061 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1065 * Added support for Kernel TLS (KTLS).
1067 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1069 * Support for RFC 5746 secure renegotiation is now required by default for
1070 SSL or TLS connections to succeed.
1074 * The signature of the `copy` functional parameter of the
1075 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1076 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1077 the signature of the `pub_decode` functional parameter of the
1078 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1079 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1083 * The error return values from some control calls (ctrl) have changed.
1087 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1091 * Many functions in the EVP_ namespace that are getters of values from
1092 implementations or contexts were renamed to include get or get0 in their
1093 names. Old names are provided as macro aliases for compatibility and
1098 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1099 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1100 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1105 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1108 * The output from the command line applications may have minor
1113 * The output from numerous "printing" may have minor changes.
1117 * Windows thread synchronization uses read/write primitives (SRWLock) when
1118 supported by the OS, otherwise CriticalSection continues to be used.
1122 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1123 work on read only BIO source/sinks that do not support these functions.
1124 This allows piping or redirection of a file BIO using stdin to be buffered
1125 into memory. This is used internally in OSSL_DECODER_from_bio().
1129 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1130 this function would return one of the values OSSL_STORE_INFO_NAME,
1131 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1132 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1133 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1134 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1135 using this function should be amended to handle the changed return value.
1139 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1140 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1141 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1142 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1143 contains more than one certificate identifier: This means that all
1144 certificates referenced there MUST be part of the validation chain.
1148 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1149 RC5, DESX and DES have been moved to the legacy provider.
1153 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1154 RIPEMD-160 have been moved to the legacy provider.
1158 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1163 * The deprecated functions EVP_PKEY_get0_RSA(),
1164 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1165 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1166 well as the similarly named "get1" functions behave differently in
1171 * A number of functions handling low-level keys or engines were deprecated
1172 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1173 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1174 EVP_PKEY_get0_siphash().
1178 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1179 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1180 will need to load the legacy crypto provider. This includes these PBE
1181 algorithms which use this KDF:
1182 - NID_pbeWithMD2AndDES_CBC
1183 - NID_pbeWithMD5AndDES_CBC
1184 - NID_pbeWithSHA1AndRC2_CBC
1185 - NID_pbeWithMD2AndRC2_CBC
1186 - NID_pbeWithMD5AndRC2_CBC
1187 - NID_pbeWithSHA1AndDES_CBC
1191 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1192 BIO_debug_callback() functions.
1196 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1197 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1201 * The RAND_METHOD APIs have been deprecated.
1205 * The SRP APIs have been deprecated.
1209 * Add a compile time option to prevent the caching of provider fetched
1210 algorithms. This is enabled by including the no-cached-fetch option
1211 at configuration time.
1215 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1216 count of PKCS12_DEFAULT_ITER.
1218 *Tomáš Mráz and Sahana Prasad*
1220 * The openssl speed command does not use low-level API calls anymore.
1224 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1227 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1229 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1233 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1234 providers may supply their own group implementations (using either the "key
1235 exchange" or the "key encapsulation" methods) which will automatically be
1236 detected and used by libssl.
1238 *Matt Caswell, Nicola Tuveri*
1240 * The undocumented function X509_certificate_type() has been deprecated;
1244 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1248 * Removed RSA padding mode for SSLv23 (which was only used for
1249 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1250 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1255 * Deprecated the obsolete X9.31 RSA key generation related functions.
1257 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1258 is not allowed to return a value > 1, this is no more taken as failure.
1260 *Viktor Dukhovni and David von Oheimb*
1262 * Deprecated the obsolete X9.31 RSA key generation related functions
1263 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1264 BN_X931_generate_prime_ex().
1268 * The default key generation method for the regular 2-prime RSA keys was
1269 changed to the FIPS 186-4 B.3.6 method.
1273 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1277 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1281 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1282 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1284 *Rich Salz, Richard Levitte, and David von Oheimb*
1286 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1290 * Deprecated `OCSP_parse_url()`.
1294 * Validation of SM2 keys has been separated from the validation of regular EC
1299 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1300 switches: a validation failure triggers an early exit, returning a failure
1301 exit status to the parent process.
1305 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1306 to ignore unknown ciphers.
1310 * The `-cipher-commands` and `-digest-commands` options
1311 of the command line utility `list` have been deprecated.
1312 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1316 * Added convenience functions for generating asymmetric key pairs:
1317 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1318 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1322 * All of the low-level EC_KEY functions have been deprecated.
1324 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1326 * Deprecated all the libcrypto and libssl error string loading
1331 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1332 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1337 * The `-crypt` option to the `passwd` command line tool has been removed.
1341 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1346 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1350 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1351 EVP_PKEY_get1_tls_encodedpoint().
1355 * The security callback, which can be customised by application code, supports
1356 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1357 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1361 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1362 interface. Their functionality remains unchanged.
1366 * Added new option for 'openssl list', '-providers', which will display the
1367 list of loaded providers, their names, version and status. It optionally
1368 displays their gettable parameters.
1372 * Removed EVP_PKEY_set_alias_type().
1376 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1377 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1381 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1382 parameters are still checked for type safety at compile time via helper
1387 * Remove the RAND_DRBG API
1389 *Paul Dale and Matthias St. Pierre*
1391 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1392 as well as actual hostnames.
1396 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1397 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1398 conversely, silently ignore DTLS protocol version bounds when configuring
1399 TLS-based contexts. The commands can be repeated to set bounds of both
1400 types. The same applies with the corresponding "min_protocol" and
1401 "max_protocol" command-line switches, in case some application uses both TLS
1404 SSL_CTX instances that are created for a fixed protocol version (e.g.
1405 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1406 attempts to apply bounds to these protocol versions would result in an
1407 error. Now only the "version-flexible" SSL_CTX instances are subject to
1408 limits in configuration files in command-line options.
1412 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1417 * Reworked the recorded ERR codes to make better space for system errors.
1418 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1419 given code is a system error (true) or an OpenSSL error (false).
1423 * Reworked the test perl framework to better allow parallel testing.
1425 *Nicola Tuveri and David von Oheimb*
1427 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1428 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1432 * 'Configure' has been changed to figure out the configuration target if
1433 none is given on the command line. Consequently, the 'config' script is
1434 now only a mere wrapper. All documentation is changed to only mention
1437 *Rich Salz and Richard Levitte*
1439 * Added a library context `OSSL_LIB_CTX` that applications as well as
1440 other libraries can use to form a separate context within which
1441 libcrypto operations are performed.
1445 * Added various `_ex` functions to the OpenSSL API that support using
1446 a non-default `OSSL_LIB_CTX`.
1450 * Handshake now fails if Extended Master Secret extension is dropped
1455 * Dropped interactive mode from the `openssl` program.
1459 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1461 *David von Oheimb and Shane Lontis*
1463 * Deprecated `EC_METHOD_get_field_type()`.
1467 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1468 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1469 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1473 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1477 * Add CAdES-BES signature verification support, mostly derived
1478 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1480 *Filipe Raimundo da Silva*
1482 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1486 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1487 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1491 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1495 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1496 EC_KEY_precompute_mult().
1500 * Deprecated EC_POINTs_mul().
1504 * Removed FIPS_mode() and FIPS_mode_set().
1508 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1512 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1513 EC_POINT_get_Jprojective_coordinates_GFp().
1517 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1518 arrays to be more easily constructed via a series of utility functions.
1519 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1520 the various push functions and finally convert to a passable OSSL_PARAM
1521 array using OSSL_PARAM_BLD_to_param().
1525 * The security strength of SHA1 and MD5 based signatures in TLS has been
1530 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1531 contain a provider side internal key.
1535 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1539 * Project text documents not yet having a proper file name extension
1540 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1541 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1545 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1546 have been converted to Markdown with the goal to produce documents
1547 which not only look pretty when viewed online in the browser, but
1548 remain well readable inside a plain text editor.
1550 To achieve this goal, a 'minimalistic' Markdown style has been applied
1551 which avoids formatting elements that interfere too much with the
1552 reading flow in the text file. For example, it
1554 * avoids [ATX headings][] and uses [setext headings][] instead
1555 (which works for `<h1>` and `<h2>` headings only).
1556 * avoids [inline links][] and uses [reference links][] instead.
1557 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1559 [ATX headings]: https://github.github.com/gfm/#atx-headings
1560 [setext headings]: https://github.github.com/gfm/#setext-headings
1561 [inline links]: https://github.github.com/gfm/#inline-link
1562 [reference links]: https://github.github.com/gfm/#reference-link
1563 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1564 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1566 *Matthias St. Pierre*
1568 * The test suite is changed to preserve results of each test recipe.
1569 A new directory test-runs/ with subdirectories named like the
1570 test recipes are created in the build tree for this purpose.
1574 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1575 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1576 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1578 *David von Oheimb, Martin Peylo*
1580 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1581 It supports arbitrary request and response content types, GET redirection,
1582 TLS, connections via HTTP(S) proxies, connections and exchange via
1583 user-defined BIOs (allowing implicit connections), persistent connections,
1584 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1585 The legacy OCSP-focused (and only partly documented) API
1586 is retained for backward compatibility, while most of it is deprecated.
1590 * Added `util/check-format.pl`, a tool for checking adherence to the
1591 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1592 The checks performed are incomplete and yield some false positives.
1593 Still the tool should be useful for detecting most typical glitches.
1597 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1598 If domain name resolution yields multiple IP addresses all of them are tried
1599 after `connect()` failures.
1603 * All of the low-level RSA functions have been deprecated.
1607 * X509 certificates signed using SHA1 are no longer allowed at security
1612 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1613 modified to use PKEY APIs. These commands are now in maintenance mode
1614 and no new features will be added to them.
1618 * The command line utility rsautl has been deprecated.
1622 * The command line utilities genrsa and rsa have been modified to use PKEY
1623 APIs. They now write PKCS#8 keys by default. These commands are now in
1624 maintenance mode and no new features will be added to them.
1628 * All of the low-level DH functions have been deprecated.
1630 *Paul Dale and Matt Caswell*
1632 * All of the low-level DSA functions have been deprecated.
1636 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1637 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1641 * Deprecated low-level ECDH and ECDSA functions.
1645 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1649 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1650 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1651 a new formulation to include all the things it can be used for,
1652 as well as words of caution.
1656 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1660 * All of the low-level HMAC functions have been deprecated.
1662 *Paul Dale and David von Oheimb*
1664 * Over two thousand fixes were made to the documentation, including:
1665 - Common options (such as -rand/-writerand, TLS version control, etc)
1666 were refactored and point to newly-enhanced descriptions in openssl.pod.
1667 - Added style conformance for all options (with help from Richard Levitte),
1668 documented all reported missing options, added a CI build to check
1669 that all options are documented and that no unimplemented options
1671 - Documented some internals, such as all use of environment variables.
1672 - Addressed all internal broken L<> references.
1676 * All of the low-level CMAC functions have been deprecated.
1680 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1681 functions have been deprecated.
1683 *Paul Dale and David von Oheimb*
1685 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1686 set of functions. The documentation mentioned negative values for some
1687 errors, but this was never the case, so the mention of negative values
1690 Code that followed the documentation and thereby check with something
1691 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1695 * All of the low-level cipher functions have been deprecated.
1697 *Matt Caswell and Paul Dale*
1699 * Removed include/openssl/opensslconf.h.in and replaced it with
1700 include/openssl/configuration.h.in, which differs in not including
1701 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1702 was added to include both.
1704 This allows internal hacks where one might need to modify the set
1705 of configured macros, for example this if deprecated symbols are
1706 still supposed to be available internally:
1708 #include <openssl/configuration.h>
1710 #undef OPENSSL_NO_DEPRECATED
1711 #define OPENSSL_SUPPRESS_DEPRECATED
1713 #include <openssl/macros.h>
1715 This should not be used by applications that use the exported
1716 symbols, as that will lead to linking errors.
1720 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1721 used in exponentiation with 512-bit moduli. No EC algorithms are
1722 affected. Analysis suggests that attacks against 2-prime RSA1024,
1723 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1724 difficult to perform and are not believed likely. Attacks against DH512
1725 are considered just feasible. However, for an attack the target would
1726 have to re-use the DH512 private key, which is not recommended anyway.
1727 Also applications directly using the low-level API BN_mod_exp may be
1728 affected if they use BN_FLG_CONSTTIME.
1733 * Most memory-debug features have been deprecated, and the functionality
1734 replaced with no-ops.
1738 * Added documentation for the STACK API.
1742 * Introduced a new method type and API, OSSL_ENCODER, to represent
1743 generic encoders. These do the same sort of job that PEM writers
1744 and d2i functions do, but with support for methods supplied by
1745 providers, and the possibility for providers to support other
1750 * Introduced a new method type and API, OSSL_DECODER, to represent
1751 generic decoders. These do the same sort of job that PEM readers
1752 and i2d functions do, but with support for methods supplied by
1753 providers, and the possibility for providers to support other
1758 * Added a .pragma directive to the syntax of configuration files, to
1759 allow varying behavior in a supported and predictable manner.
1760 Currently added pragma:
1764 This allows dollar signs to be a keyword character unless it's
1765 followed by a opening brace or parenthesis. This is useful for
1766 platforms where dollar signs are commonly used in names, such as
1767 volume names and system directory names on VMS.
1771 * Added functionality to create an EVP_PKEY from user data.
1775 * Change the interpretation of the '--api' configuration option to
1776 mean that this is a desired API compatibility level with no
1777 further meaning. The previous interpretation, that this would
1778 also mean to remove all deprecated symbols up to and including
1779 the given version, no requires that 'no-deprecated' is also used
1780 in the configuration.
1782 When building applications, the desired API compatibility level
1783 can be set with the OPENSSL_API_COMPAT macro like before. For
1784 API compatibility version below 3.0, the old style numerical
1785 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1786 For version 3.0 and on, the value is expected to be the decimal
1787 value calculated from the major and minor version like this:
1789 MAJOR * 10000 + MINOR * 100
1793 -DOPENSSL_API_COMPAT=30000 For 3.0
1794 -DOPENSSL_API_COMPAT=30200 For 3.2
1796 To hide declarations that are deprecated up to and including the
1797 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1798 given when building the application as well.
1802 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1803 access to certificate and CRL stores via URIs and OSSL_STORE
1806 This adds the following functions:
1808 - X509_LOOKUP_store()
1809 - X509_STORE_load_file()
1810 - X509_STORE_load_path()
1811 - X509_STORE_load_store()
1812 - SSL_add_store_cert_subjects_to_stack()
1813 - SSL_CTX_set_default_verify_store()
1814 - SSL_CTX_load_verify_file()
1815 - SSL_CTX_load_verify_dir()
1816 - SSL_CTX_load_verify_store()
1820 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1821 The presence of this system service is determined at run-time.
1825 * Added functionality to create an EVP_PKEY context based on data
1826 for methods from providers. This takes an algorithm name and a
1827 property query string and simply stores them, with the intent
1828 that any operation that uses this context will use those strings
1829 to fetch the needed methods implicitly, thereby making the port
1830 of application written for pre-3.0 OpenSSL easier.
1834 * The undocumented function NCONF_WIN32() has been deprecated; for
1835 conversion details see the HISTORY section of doc/man5/config.pod
1839 * Introduced the new functions EVP_DigestSignInit_ex() and
1840 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1841 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1842 pages for further details.
1846 * Over two thousand fixes were made to the documentation, including:
1847 adding missing command flags, better style conformance, documentation
1850 *Rich Salz, Richard Levitte*
1852 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1853 X25519, X448, Ed25519 and Ed448.
1857 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1862 * Deprecated the public definition of `ERR_STATE` as well as the function
1863 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1868 * Added ERR functionality to give callers access to the stored function
1869 names that have replaced the older function code based functions.
1871 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1872 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1873 ERR_peek_error_all() and ERR_peek_last_error_all().
1875 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1876 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1877 ERR_func_error_string().
1881 * Extended testing to be verbose for failing tests only. The make variables
1882 VERBOSE_FAILURE or VF can be used to enable this:
1884 $ make VF=1 test # Unix
1885 $ mms /macro=(VF=1) test ! OpenVMS
1886 $ nmake VF=1 test # Windows
1890 * Added the `-copy_extensions` option to the `x509` command for use with
1891 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1892 all extensions in the request are copied to the certificate or vice versa.
1894 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1896 * Added the `-copy_extensions` option to the `req` command for use with
1897 `-x509`. When given with the `copy` or `copyall` argument,
1898 all extensions in the certification request are copied to the certificate.
1902 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1903 they generate are by default RFC 5280 compliant in the following sense:
1904 There is a subjectKeyIdentifier extension with a hash value of the public key
1905 and for not self-signed certs there is an authorityKeyIdentifier extension
1906 with a keyIdentifier field or issuer information identifying the signing key.
1907 This is done unless some configuration overrides the new default behavior,
1908 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1912 * Added several checks to `X509_verify_cert()` according to requirements in
1913 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1914 (which may be done by using the CLI option `-x509_strict`):
1915 * The basicConstraints of CA certificates must be marked critical.
1916 * CA certificates must explicitly include the keyUsage extension.
1917 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1918 * The issuer name of any certificate must not be empty.
1919 * The subject name of CA certs, certs with keyUsage crlSign,
1920 and certs without subjectAlternativeName must not be empty.
1921 * If a subjectAlternativeName extension is given it must not be empty.
1922 * The signatureAlgorithm field and the cert signature must be consistent.
1923 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1924 must not be marked critical.
1925 * The authorityKeyIdentifier must be given for X.509v3 certs
1926 unless they are self-signed.
1927 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1931 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1932 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1936 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1937 used even when parsing explicit parameters, when loading a encoded key
1938 or calling `EC_GROUP_new_from_ecpkparameters()`/
1939 `EC_GROUP_new_from_ecparameters()`.
1940 This prevents bypass of security hardening and performance gains,
1941 especially for curves with specialized EC_METHODs.
1942 By default, if a key encoded with explicit parameters is loaded and later
1943 encoded, the output is still encoded with explicit parameters, even if
1944 internally a "named" EC_GROUP is used for computation.
1948 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1949 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1950 NULL. After this change, only the cofactor parameter can be NULL. It also
1951 does some minimal sanity checks on the passed order.
1956 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1957 An attack is simple, if the first CMS_recipientInfo is valid but the
1958 second CMS_recipientInfo is chosen ciphertext. If the second
1959 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1960 encryption key will be replaced by garbage, and the message cannot be
1961 decoded, but if the RSA decryption fails, the correct encryption key is
1962 used and the recipient will not notice the attack.
1963 As a work around for this potential attack the length of the decrypted
1964 key must be equal to the cipher default key length, in case the
1965 certificate is not given and all recipientInfo are tried out.
1966 The old behaviour can be re-enabled in the CMS code by setting the
1967 CMS_DEBUG_DECRYPT flag.
1971 * Early start up entropy quality from the DEVRANDOM seed source has been
1972 improved for older Linux systems. The RAND subsystem will wait for
1973 /dev/random to be producing output before seeding from /dev/urandom.
1974 The seeded state is stored for future library initialisations using
1975 a system global shared memory segment. The shared memory identifier
1976 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1977 the desired value. The default identifier is 114.
1981 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1982 when primes for RSA keys are computed.
1983 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1984 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1985 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1986 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1987 This avoids possible fingerprinting of newly generated RSA modules.
1991 * Correct the extended master secret constant on EBCDIC systems. Without this
1992 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1993 negotiate EMS will fail. Unfortunately this also means that TLS connections
1994 between EBCDIC systems with this fix, and EBCDIC systems without this
1995 fix will fail if they negotiate EMS.
1999 * Changed the library initialisation so that the config file is now loaded
2000 by default. This was already the case for libssl. It now occurs for both
2001 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2002 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2006 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2007 where the former acts as a replacement for `ERR_put_error()`, and the
2008 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2009 `ERR_raise_data()` adds more flexibility by taking a format string and
2010 an arbitrary number of arguments following it, to be processed with
2015 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2016 to check if a named provider is loaded and available. When called, it
2017 will also activate all fallback providers if such are still present.
2021 * Enforce a minimum DH modulus size of 512 bits.
2025 * Changed DH parameters to generate the order q subgroup instead of 2q.
2026 Previously generated DH parameters are still accepted by DH_check
2027 but DH_generate_key works around that by clearing bit 0 of the
2028 private key for those. This avoids leaking bit 0 of the private key.
2032 * Significantly reduce secure memory usage by the randomness pools.
2036 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2041 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2042 algorithms. An implementation of a key exchange algorithm can be obtained
2043 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2044 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2045 the older EVP_PKEY_derive_init() function. See the man pages for the new
2046 functions for further details.
2050 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2054 * Removed the function names from error messages and deprecated the
2059 * Removed NextStep support and the macro OPENSSL_UNISTD
2063 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2064 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2065 Also removed "export var as function" capability; we do not export
2066 variables, only functions.
2070 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2071 an error and 1 indicating success. In previous versions of OpenSSL this
2072 was a void type. If a key was set longer than the maximum possible this
2077 * Support SM2 signing and verification schemes with X509 certificate.
2081 * Use SHA256 as the default digest for TS query in the `ts` app.
2085 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2089 * Default cipher lists/suites are now available via a function, the
2090 #defines are deprecated.
2094 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2095 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2096 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2100 * Join the directories crypto/x509 and crypto/x509v3
2104 * Added command 'openssl kdf' that uses the EVP_KDF API.
2108 * Added command 'openssl mac' that uses the EVP_MAC API.
2112 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2113 as default directories. Also added the command 'openssl info'
2114 for scripting purposes.
2118 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2123 * Add prediction resistance to the DRBG reseeding process.
2127 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2128 mandated by IEEE Std 1619-2018.
2132 * Added newline escaping functionality to a filename when using openssl dgst.
2133 This output format is to replicate the output format found in the `*sum`
2134 checksum programs. This aims to preserve backward compatibility.
2136 *Matt Eaton, Richard Levitte, and Paul Dale*
2138 * Removed the heartbeat message in DTLS feature, as it has very
2139 little usage and doesn't seem to fulfill a valuable purpose.
2140 The configuration option is now deprecated.
2144 * Changed the output of 'openssl {digestname} < file' to display the
2145 digest name in its output.
2149 * Added a new generic trace API which provides support for enabling
2150 instrumentation through trace output.
2152 *Richard Levitte & Matthias St. Pierre*
2154 * Added build tests for C++. These are generated files that only do one
2155 thing, to include one public OpenSSL head file each. This tests that
2156 the public header files can be usefully included in a C++ application.
2158 This test isn't enabled by default. It can be enabled with the option
2159 'enable-buildtest-c++'.
2163 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2167 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2171 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2175 * Added KMAC to EVP_MAC.
2179 * Added property based algorithm implementation selection framework to
2184 * Added SCA hardening for modular field inversion in EC_GROUP through
2185 a new dedicated field_inv() pointer in EC_METHOD.
2186 This also addresses a leakage affecting conversions from projective
2187 to affine coordinates.
2189 *Billy Bob Brumley, Nicola Tuveri*
2191 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2192 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2193 those algorithms that were already supported through the EVP_PKEY API
2194 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2195 and scrypt are now wrappers that call EVP_KDF.
2199 * Build devcrypto engine as a dynamic engine.
2201 *Eneas U de Queiroz*
2203 * Add keyed BLAKE2 to EVP_MAC.
2207 * Fix a bug in the computation of the endpoint-pair shared secret used
2208 by DTLS over SCTP. This breaks interoperability with older versions
2209 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2210 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2211 interoperability with such broken implementations. However, enabling
2212 this switch breaks interoperability with correct implementations.
2214 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2215 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2219 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2223 * Changed the license to the Apache License v2.0.
2227 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2229 - Major releases (indicated by incrementing the MAJOR release number)
2230 may introduce incompatible API/ABI changes.
2231 - Minor releases (indicated by incrementing the MINOR release number)
2232 may introduce new features but retain API/ABI compatibility.
2233 - Patch releases (indicated by incrementing the PATCH number)
2234 are intended for bug fixes and other improvements of existing
2235 features only (like improving performance or adding documentation)
2236 and retain API/ABI compatibility.
2240 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2244 * Remove the 'dist' target and add a tarball building script. The
2245 'dist' target has fallen out of use, and it shouldn't be
2246 necessary to configure just to create a source distribution.
2250 * Recreate the OS390-Unix config target. It no longer relies on a
2251 special script like it did for OpenSSL pre-1.1.0.
2255 * Instead of having the source directories listed in Configure, add
2256 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2261 * Add GMAC to EVP_MAC.
2265 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2269 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2270 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2271 to facilitate the continued use of MACs through raw private keys in
2272 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2276 * Deprecate ECDH_KDF_X9_62().
2280 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2281 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2282 are retained for backwards compatibility.
2286 * AES-XTS mode now enforces that its two keys are different to mitigate
2287 the attacked described in "Efficient Instantiations of Tweakable
2288 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2289 Details of this attack can be obtained from:
2290 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2294 * Rename the object files, i.e. give them other names than in previous
2295 versions. Their names now include the name of the final product, as
2296 well as its type mnemonic (bin, lib, shlib).
2300 * Added new option for 'openssl list', '-objects', which will display the
2301 list of built in objects, i.e. OIDs with names.
2305 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2306 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2311 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2312 improves application performance by removing data copies and providing
2313 applications with zero-copy system calls such as sendfile and splice.
2317 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2321 * `PKCS12_parse` now maintains the order of the parsed certificates
2322 when outputting them via `*ca` (rather than reversing it).
2326 * Deprecated pthread fork support methods.
2330 * Added support for FFDHE key exchange in TLS 1.3.
2334 * Added a new concept for OpenSSL plugability: providers. This
2335 functionality is designed to replace the ENGINE API and ENGINE
2336 implementations, and to be much more dynamic, allowing provider
2337 authors to introduce new algorithms among other things, as long as
2338 there's an API that supports the algorithm type.
2340 With this concept comes a new core API for interaction between
2341 libcrypto and provider implementations. Public libcrypto functions
2342 that want to use providers do so through this core API.
2344 The main documentation for this core API is found in
2345 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2346 refer to other manuals describing the API specific for supported
2347 algorithm types (also called operations).
2354 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2356 * Avoid loading of a dynamic engine twice.
2360 * Prioritise DANE TLSA issuer certs over peer certs
2364 * Fixed random API for MacOS prior to 10.12
2366 These MacOS versions don't support the CommonCrypto APIs
2370 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2372 * Fixed an SM2 Decryption Buffer Overflow.
2374 In order to decrypt SM2 encrypted data an application is expected to
2375 call the API function EVP_PKEY_decrypt(). Typically an application will
2376 call this function twice. The first time, on entry, the "out" parameter
2377 can be NULL and, on exit, the "outlen" parameter is populated with the
2378 buffer size required to hold the decrypted plaintext. The application
2379 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2380 again, but this time passing a non-NULL value for the "out" parameter.
2382 A bug in the implementation of the SM2 decryption code means that the
2383 calculation of the buffer size required to hold the plaintext returned
2384 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2385 size required by the second call. This can lead to a buffer overflow
2386 when EVP_PKEY_decrypt() is called by the application a second time with
2387 a buffer that is too small.
2389 A malicious attacker who is able present SM2 content for decryption to
2390 an application could cause attacker chosen data to overflow the buffer
2391 by up to a maximum of 62 bytes altering the contents of other data held
2392 after the buffer, possibly changing application behaviour or causing
2393 the application to crash. The location of the buffer is application
2394 dependent but is typically heap allocated.
2399 * Fixed various read buffer overruns processing ASN.1 strings
2401 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2402 structure which contains a buffer holding the string data and a field
2403 holding the buffer length. This contrasts with normal C strings which
2404 are represented as a buffer for the string data which is terminated
2405 with a NUL (0) byte.
2407 Although not a strict requirement, ASN.1 strings that are parsed using
2408 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2409 well as any string whose value has been set with the ASN1_STRING_set()
2410 function will additionally NUL terminate the byte array in the
2411 ASN1_STRING structure.
2413 However, it is possible for applications to directly construct valid
2414 ASN1_STRING structures which do not NUL terminate the byte array by
2415 directly setting the "data" and "length" fields in the ASN1_STRING
2416 array. This can also happen by using the ASN1_STRING_set0() function.
2418 Numerous OpenSSL functions that print ASN.1 data have been found to
2419 assume that the ASN1_STRING byte array will be NUL terminated, even
2420 though this is not guaranteed for strings that have been directly
2421 constructed. Where an application requests an ASN.1 structure to be
2422 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2423 been directly constructed by the application without NUL terminating
2424 the "data" field, then a read buffer overrun can occur.
2426 The same thing can also occur during name constraints processing
2427 of certificates (for example if a certificate has been directly
2428 constructed by the application instead of loading it via the OpenSSL
2429 parsing functions, and the certificate contains non NUL terminated
2430 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2431 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2433 If a malicious actor can cause an application to directly construct an
2434 ASN1_STRING and then process it through one of the affected OpenSSL
2435 functions then this issue could be hit. This might result in a crash
2436 (causing a Denial of Service attack). It could also result in the
2437 disclosure of private memory contents (such as private keys, or
2438 sensitive plaintext).
2443 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2445 * Fixed a problem with verifying a certificate chain when using the
2446 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2447 the certificates present in a certificate chain. It is not set by default.
2449 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2450 the chain that have explicitly encoded elliptic curve parameters was added
2451 as an additional strict check.
2453 An error in the implementation of this check meant that the result of a
2454 previous check to confirm that certificates in the chain are valid CA
2455 certificates was overwritten. This effectively bypasses the check
2456 that non-CA certificates must not be able to issue other certificates.
2458 If a "purpose" has been configured then there is a subsequent opportunity
2459 for checks that the certificate is a valid CA. All of the named "purpose"
2460 values implemented in libcrypto perform this check. Therefore, where
2461 a purpose is set the certificate chain will still be rejected even when the
2462 strict flag has been used. A purpose is set by default in libssl client and
2463 server certificate verification routines, but it can be overridden or
2464 removed by an application.
2466 In order to be affected, an application must explicitly set the
2467 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2468 for the certificate verification or, in the case of TLS client or server
2469 applications, override the default purpose.
2474 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2475 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2476 renegotiation ClientHello omits the signature_algorithms extension (where it
2477 was present in the initial ClientHello), but includes a
2478 signature_algorithms_cert extension then a NULL pointer dereference will
2479 result, leading to a crash and a denial of service attack.
2481 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2482 (which is the default configuration). OpenSSL TLS clients are not impacted by
2486 *Peter Kästle and Samuel Sapalski*
2488 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2490 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2491 create a unique hash value based on the issuer and serial number data
2492 contained within an X509 certificate. However, it was failing to correctly
2493 handle any errors that may occur while parsing the issuer field (which might
2494 occur if the issuer field is maliciously constructed). This may subsequently
2495 result in a NULL pointer deref and a crash leading to a potential denial of
2501 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2502 padding mode to correctly check for rollback attacks. This is considered a
2503 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2508 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2509 functions. Previously they could overflow the output length argument in some
2510 cases where the input length is close to the maximum permissible length for
2511 an integer on the platform. In such cases the return value from the function
2512 call would be 1 (indicating success), but the output length value would be
2513 negative. This could cause applications to behave incorrectly or crash.
2518 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2519 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2520 could be exploited in a side channel attack to recover the password. Since
2521 the attack is local host only this is outside of the current OpenSSL
2522 threat model and therefore no CVE is assigned.
2524 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2529 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2531 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2532 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2533 If an attacker can control both items being compared then this could lead
2534 to a possible denial of service attack. OpenSSL itself uses the
2535 GENERAL_NAME_cmp function for two purposes:
2536 1) Comparing CRL distribution point names between an available CRL and a
2537 CRL distribution point embedded in an X509 certificate
2538 2) When verifying that a timestamp response token signer matches the
2539 timestamp authority name (exposed via the API functions
2540 TS_RESP_verify_response and TS_RESP_verify_token)
2545 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2547 * Certificates with explicit curve parameters are now disallowed in
2548 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2552 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2553 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2554 conversely, silently ignore DTLS protocol version bounds when configuring
2555 TLS-based contexts. The commands can be repeated to set bounds of both
2556 types. The same applies with the corresponding "min_protocol" and
2557 "max_protocol" command-line switches, in case some application uses both TLS
2560 SSL_CTX instances that are created for a fixed protocol version (e.g.
2561 TLSv1_server_method()) also silently ignore version bounds. Previously
2562 attempts to apply bounds to these protocol versions would result in an
2563 error. Now only the "version-flexible" SSL_CTX instances are subject to
2564 limits in configuration files in command-line options.
2568 * Handshake now fails if Extended Master Secret extension is dropped
2573 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2575 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2577 * Fixed segmentation fault in SSL_check_chain()
2578 Server or client applications that call the SSL_check_chain() function
2579 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2580 dereference as a result of incorrect handling of the
2581 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2582 or unrecognised signature algorithm is received from the peer. This could
2583 be exploited by a malicious peer in a Denial of Service attack.
2588 * Added AES consttime code for no-asm configurations
2589 an optional constant time support for AES was added
2590 when building openssl for no-asm.
2591 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2592 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2593 At this time this feature is by default disabled.
2594 It will be enabled by default in 3.0.
2598 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2600 * Revert the change of EOF detection while reading in libssl to avoid
2601 regressions in applications depending on the current way of reporting
2602 the EOF. As the existing method is not fully accurate the change to
2603 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2604 branch and will be present in the 3.0 release.
2608 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2609 when primes for RSA keys are computed.
2610 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2611 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2612 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2613 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2614 This avoids possible fingerprinting of newly generated RSA modules.
2618 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2620 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2621 while reading in libssl then we would report an error back to the
2622 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2623 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2624 therefore give a hint as to what went wrong.
2628 * Check that ed25519 and ed448 are allowed by the security level. Previously
2629 signature algorithms not using an MD were not being checked that they were
2630 allowed by the security level.
2634 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2635 was not quite right. The behaviour was not consistent between resumption
2636 and normal handshakes, and also not quite consistent with historical
2637 behaviour. The behaviour in various scenarios has been clarified and
2638 it has been updated to make it match historical behaviour as closely as
2643 * *[VMS only]* The header files that the VMS compilers include automatically,
2644 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2645 that the C++ compiler doesn't understand. This is a shortcoming in the
2646 compiler, but can be worked around with `__cplusplus` guards.
2648 C++ applications that use OpenSSL libraries must be compiled using the
2649 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2650 functions. Otherwise, only functions with symbols of less than 31
2651 characters can be used, as the linker will not be able to successfully
2652 resolve symbols with longer names.
2656 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2657 The presence of this system service is determined at run-time.
2661 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2666 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2668 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2669 number generator (RNG). This was intended to include protection in the
2670 event of a fork() system call in order to ensure that the parent and child
2671 processes did not share the same RNG state. However, this protection was not
2672 being used in the default case.
2674 A partial mitigation for this issue is that the output from a high
2675 precision timer is mixed into the RNG state so the likelihood of a parent
2676 and child process sharing state is significantly reduced.
2678 If an application already calls OPENSSL_init_crypto() explicitly using
2679 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2682 *Matthias St. Pierre*
2684 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2685 used even when parsing explicit parameters, when loading a encoded key
2686 or calling `EC_GROUP_new_from_ecpkparameters()`/
2687 `EC_GROUP_new_from_ecparameters()`.
2688 This prevents bypass of security hardening and performance gains,
2689 especially for curves with specialized EC_METHODs.
2690 By default, if a key encoded with explicit parameters is loaded and later
2691 encoded, the output is still encoded with explicit parameters, even if
2692 internally a "named" EC_GROUP is used for computation.
2696 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2697 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2698 NULL. After this change, only the cofactor parameter can be NULL. It also
2699 does some minimal sanity checks on the passed order.
2704 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2705 An attack is simple, if the first CMS_recipientInfo is valid but the
2706 second CMS_recipientInfo is chosen ciphertext. If the second
2707 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2708 encryption key will be replaced by garbage, and the message cannot be
2709 decoded, but if the RSA decryption fails, the correct encryption key is
2710 used and the recipient will not notice the attack.
2711 As a work around for this potential attack the length of the decrypted
2712 key must be equal to the cipher default key length, in case the
2713 certificate is not given and all recipientInfo are tried out.
2714 The old behaviour can be re-enabled in the CMS code by setting the
2715 CMS_DEBUG_DECRYPT flag.
2720 * Early start up entropy quality from the DEVRANDOM seed source has been
2721 improved for older Linux systems. The RAND subsystem will wait for
2722 /dev/random to be producing output before seeding from /dev/urandom.
2723 The seeded state is stored for future library initialisations using
2724 a system global shared memory segment. The shared memory identifier
2725 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2726 the desired value. The default identifier is 114.
2730 * Correct the extended master secret constant on EBCDIC systems. Without this
2731 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2732 negotiate EMS will fail. Unfortunately this also means that TLS connections
2733 between EBCDIC systems with this fix, and EBCDIC systems without this
2734 fix will fail if they negotiate EMS.
2738 * Use Windows installation paths in the mingw builds
2740 Mingw isn't a POSIX environment per se, which means that Windows
2741 paths should be used for installation.
2746 * Changed DH_check to accept parameters with order q and 2q subgroups.
2747 With order 2q subgroups the bit 0 of the private key is not secret
2748 but DH_generate_key works around that by clearing bit 0 of the
2749 private key for those. This avoids leaking bit 0 of the private key.
2753 * Significantly reduce secure memory usage by the randomness pools.
2757 * Revert the DEVRANDOM_WAIT feature for Linux systems
2759 The DEVRANDOM_WAIT feature added a select() call to wait for the
2760 /dev/random device to become readable before reading from the
2761 /dev/urandom device.
2763 It turned out that this change had negative side effects on
2764 performance which were not acceptable. After some discussion it
2765 was decided to revert this feature and leave it up to the OS
2766 resp. the platform maintainer to ensure a proper initialization
2767 during early boot time.
2769 *Matthias St. Pierre*
2771 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2773 * Add build tests for C++. These are generated files that only do one
2774 thing, to include one public OpenSSL head file each. This tests that
2775 the public header files can be usefully included in a C++ application.
2777 This test isn't enabled by default. It can be enabled with the option
2778 'enable-buildtest-c++'.
2782 * Enable SHA3 pre-hashing for ECDSA and DSA.
2786 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2787 This changes the size when using the `genpkey` command when no size is given.
2788 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2789 generation commands to use 2048 bits by default.
2793 * Reorganize the manual pages to consistently have RETURN VALUES,
2794 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2795 util/fix-doc-nits accordingly.
2797 *Paul Yang, Joshua Lock*
2799 * Add the missing accessor EVP_PKEY_get0_engine()
2803 * Have commands like `s_client` and `s_server` output the signature scheme
2804 along with other cipher suite parameters when debugging.
2808 * Make OPENSSL_config() error agnostic again.
2812 * Do the error handling in RSA decryption constant time.
2816 * Prevent over long nonces in ChaCha20-Poly1305.
2818 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2819 for every encryption operation. RFC 7539 specifies that the nonce value
2820 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2821 and front pads the nonce with 0 bytes if it is less than 12
2822 bytes. However it also incorrectly allows a nonce to be set of up to 16
2823 bytes. In this case only the last 12 bytes are significant and any
2824 additional leading bytes are ignored.
2826 It is a requirement of using this cipher that nonce values are
2827 unique. Messages encrypted using a reused nonce value are susceptible to
2828 serious confidentiality and integrity attacks. If an application changes
2829 the default nonce length to be longer than 12 bytes and then makes a
2830 change to the leading bytes of the nonce expecting the new value to be a
2831 new unique nonce then such an application could inadvertently encrypt
2832 messages with a reused nonce.
2834 Additionally the ignored bytes in a long nonce are not covered by the
2835 integrity guarantee of this cipher. Any application that relies on the
2836 integrity of these ignored leading bytes of a long nonce may be further
2837 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2838 is safe because no such use sets such a long nonce value. However user
2839 applications that use this cipher directly and set a non-default nonce
2840 length to be longer than 12 bytes may be vulnerable.
2842 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2848 * Add DEVRANDOM_WAIT feature for Linux systems
2850 On older Linux systems where the getrandom() system call is not available,
2851 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2852 Contrary to getrandom(), the /dev/urandom device will not block during
2853 early boot when the kernel CSPRNG has not been seeded yet.
2855 To mitigate this known weakness, use select() to wait for /dev/random to
2856 become readable before reading from /dev/urandom.
2858 * Ensure that SM2 only uses SM3 as digest algorithm
2862 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2864 * Change the info callback signals for the start and end of a post-handshake
2865 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2866 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2867 confused by this and assume that a TLSv1.2 renegotiation has started. This
2868 can break KeyUpdate handling. Instead we no longer signal the start and end
2869 of a post handshake message exchange (although the messages themselves are
2870 still signalled). This could break some applications that were expecting
2871 the old signals. However without this KeyUpdate is not usable for many
2876 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2878 * Timing vulnerability in DSA signature generation
2880 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2881 timing side channel attack. An attacker could use variations in the signing
2882 algorithm to recover the private key.
2884 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2889 * Timing vulnerability in ECDSA signature generation
2891 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2892 timing side channel attack. An attacker could use variations in the signing
2893 algorithm to recover the private key.
2895 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2900 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2901 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2902 of two gigabytes and the error handling improved.
2904 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2905 categorized as a normal bug, not a security issue, because the DRBG reseeds
2906 automatically and is fully functional even without additional randomness
2907 provided by the application.
2909 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2911 * Add a new ClientHello callback. Provides a callback interface that gives
2912 the application the ability to adjust the nascent SSL object at the
2913 earliest stage of ClientHello processing, immediately after extensions have
2914 been collected but before they have been processed. In particular, this
2915 callback can adjust the supported TLS versions in response to the contents
2920 * Add SM2 base algorithm support.
2924 * s390x assembly pack: add (improved) hardware-support for the following
2925 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2926 aes-cfb/cfb8, aes-ecb.
2930 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2931 parameter is no longer accepted, as it leads to a corrupt table. NULL
2932 pem_str is reserved for alias entries only.
2936 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2937 step for prime curves. The new implementation is based on formulae from
2938 differential addition-and-doubling in homogeneous projective coordinates
2939 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2940 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2941 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2942 to work in projective coordinates.
2944 *Billy Bob Brumley, Nicola Tuveri*
2946 * Change generating and checking of primes so that the error rate of not
2947 being prime depends on the intended use based on the size of the input.
2948 For larger primes this will result in more rounds of Miller-Rabin.
2949 The maximal error rate for primes with more than 1080 bits is lowered
2952 *Kurt Roeckx, Annie Yousar*
2954 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2958 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2959 moving between systems, and to avoid confusion when a Windows build is
2960 done with mingw vs with MSVC. For POSIX installs, there's still a
2961 symlink or copy named 'tsget' to avoid that confusion as well.
2965 * Revert blinding in ECDSA sign and instead make problematic addition
2966 length-invariant. Switch even to fixed-length Montgomery multiplication.
2970 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2971 step for binary curves. The new implementation is based on formulae from
2972 differential addition-and-doubling in mixed Lopez-Dahab projective
2973 coordinates, modified to independently blind the operands.
2975 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2977 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2978 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2979 EC_METHODs to implement their own specialized "ladder step", to take
2980 advantage of more favorable coordinate systems or more efficient
2981 differential addition-and-doubling algorithms.
2983 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2985 * Modified the random device based seed sources to keep the relevant
2986 file descriptors open rather than reopening them on each access.
2987 This allows such sources to operate in a chroot() jail without
2988 the associated device nodes being available. This behaviour can be
2989 controlled using RAND_keep_random_devices_open().
2993 * Numerous side-channel attack mitigations have been applied. This may have
2994 performance impacts for some algorithms for the benefit of improved
2995 security. Specific changes are noted in this change log by their respective
3000 * AIX shared library support overhaul. Switch to AIX "natural" way of
3001 handling shared libraries, which means collecting shared objects of
3002 different versions and bitnesses in one common archive. This allows to
3003 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3004 doesn't affect the way 3rd party applications are linked, only how
3005 multi-version installation is managed.
3009 * Make ec_group_do_inverse_ord() more robust and available to other
3010 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3011 mitigations are applied to the fallback BN_mod_inverse().
3012 When using this function rather than BN_mod_inverse() directly, new
3013 EC cryptosystem implementations are then safer-by-default.
3017 * Add coordinate blinding for EC_POINT and implement projective
3018 coordinate blinding for generic prime curves as a countermeasure to
3019 chosen point SCA attacks.
3021 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3023 * Add blinding to ECDSA and DSA signatures to protect against side channel
3024 attacks discovered by Keegan Ryan (NCC Group).
3028 * Enforce checking in the `pkeyutl` command to ensure that the input
3029 length does not exceed the maximum supported digest length when performing
3030 a sign, verify or verifyrecover operation.
3034 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3035 I/O in combination with something like select() or poll() will hang. This
3036 can be turned off again using SSL_CTX_clear_mode().
3037 Many applications do not properly handle non-application data records, and
3038 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3039 around the problems in those applications, but can also break some.
3040 It's recommended to read the manpages about SSL_read(), SSL_write(),
3041 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3042 SSL_CTX_set_read_ahead() again.
3046 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3047 now allow empty (zero character) pass phrases.
3051 * Apply blinding to binary field modular inversion and remove patent
3052 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3056 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3057 binary and prime elliptic curves.
3061 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3062 constant time fixed point multiplication.
3066 * Revise elliptic curve scalar multiplication with timing attack
3067 defenses: ec_wNAF_mul redirects to a constant time implementation
3068 when computing fixed point and variable point multiplication (which
3069 in OpenSSL are mostly used with secret scalars in keygen, sign,
3070 ECDH derive operations).
3071 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3074 * Updated CONTRIBUTING
3078 * Updated DRBG / RAND to request nonce and additional low entropy
3079 randomness from the system.
3081 *Matthias St. Pierre*
3083 * Updated 'openssl rehash' to use OpenSSL consistent default.
3087 * Moved the load of the ssl_conf module to libcrypto, which helps
3088 loading engines that libssl uses before libssl is initialised.
3092 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3096 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3098 *Ingo Schwarze, Rich Salz*
3100 * Added output of accepting IP address and port for 'openssl s_server'
3104 * Added a new API for TLSv1.3 ciphersuites:
3105 SSL_CTX_set_ciphersuites()
3106 SSL_set_ciphersuites()
3110 * Memory allocation failures consistently add an error to the error
3115 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3116 in libcrypto when run as setuid/setgid.
3120 * Load any config file by default when libssl is used.
3124 * Added new public header file <openssl/rand_drbg.h> and documentation
3125 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3127 *Matthias St. Pierre*
3129 * QNX support removed (cannot find contributors to get their approval
3130 for the license change).
3134 * TLSv1.3 replay protection for early data has been implemented. See the
3135 SSL_read_early_data() man page for further details.
3139 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3140 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3141 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3142 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3143 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3144 configuration has been separated out. See the ciphers man page or the
3145 SSL_CTX_set_ciphersuites() man page for more information.
3149 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3150 in responder mode now supports the new "-multi" option, which
3151 spawns the specified number of child processes to handle OCSP
3152 requests. The "-timeout" option now also limits the OCSP
3153 responder's patience to wait to receive the full client request
3154 on a newly accepted connection. Child processes are respawned
3155 as needed, and the CA index file is automatically reloaded
3156 when changed. This makes it possible to run the "ocsp" responder
3157 as a long-running service, making the OpenSSL CA somewhat more
3158 feature-complete. In this mode, most diagnostic messages logged
3159 after entering the event loop are logged via syslog(3) rather than
3164 * Added support for X448 and Ed448. Heavily based on original work by
3169 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3170 objects loaded. This adds the functions OSSL_STORE_expect() and
3171 OSSL_STORE_find() as well as needed tools to construct searches and
3172 get the search data out of them.
3176 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3177 version of OpenSSL should review their configuration settings to ensure
3178 that they are still appropriate for TLSv1.3. For further information see:
3179 <https://wiki.openssl.org/index.php/TLS1.3>
3183 * Grand redesign of the OpenSSL random generator
3185 The default RAND method now utilizes an AES-CTR DRBG according to
3186 NIST standard SP 800-90Ar1. The new random generator is essentially
3187 a port of the default random generator from the OpenSSL FIPS 2.0
3188 object module. It is a hybrid deterministic random bit generator
3189 using an AES-CTR bit stream and which seeds and reseeds itself
3190 automatically using trusted system entropy sources.
3192 Some of its new features are:
3193 - Support for multiple DRBG instances with seed chaining.
3194 - The default RAND method makes use of a DRBG.
3195 - There is a public and private DRBG instance.
3196 - The DRBG instances are fork-safe.
3197 - Keep all global DRBG instances on the secure heap if it is enabled.
3198 - The public and private DRBG instance are per thread for lock free
3201 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3203 * Changed Configure so it only says what it does and doesn't dump
3204 so much data. Instead, ./configdata.pm should be used as a script
3205 to display all sorts of configuration data.
3209 * Added processing of "make variables" to Configure.
3213 * Added SHA512/224 and SHA512/256 algorithm support.
3217 * The last traces of Netware support, first removed in 1.1.0, have
3222 * Get rid of Makefile.shared, and in the process, make the processing
3223 of certain files (rc.obj, or the .def/.map/.opt files produced from
3224 the ordinal files) more visible and hopefully easier to trace and
3225 debug (or make silent).
3229 * Make it possible to have environment variable assignments as
3230 arguments to config / Configure.
3234 * Add multi-prime RSA (RFC 8017) support.
3238 * Add SM3 implemented according to GB/T 32905-2016
3239 *Jack Lloyd <jack.lloyd@ribose.com>,*
3240 *Ronald Tse <ronald.tse@ribose.com>,*
3241 *Erick Borsboom <erick.borsboom@ribose.com>*
3243 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3244 as documented in RFC6066.
3245 Based on a patch from Tomasz Moń
3247 *Filipe Raimundo da Silva*
3249 * Add SM4 implemented according to GB/T 32907-2016.
3250 *Jack Lloyd <jack.lloyd@ribose.com>,*
3251 *Ronald Tse <ronald.tse@ribose.com>,*
3252 *Erick Borsboom <erick.borsboom@ribose.com>*
3254 * Reimplement -newreq-nodes and ERR_error_string_n; the
3255 original author does not agree with the license change.
3259 * Add ARIA AEAD TLS support.
3263 * Some macro definitions to support VS6 have been removed. Visual
3264 Studio 6 has not worked since 1.1.0
3268 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3269 without clearing the errors.
3273 * Add "atfork" functions. If building on a system that without
3274 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3275 requirements. The RAND facility now uses/requires this.
3283 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3284 not possible to disable entirely. However, it's still possible to
3285 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3288 To disable, configure with 'no-ui-console'. 'no-ui' is still
3289 possible to use as an alias. Check at compile time with the
3290 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3291 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3295 * Add a STORE module, which implements a uniform and URI based reader of
3296 stores that can contain keys, certificates, CRLs and numerous other
3297 objects. The main API is loosely based on a few stdio functions,
3298 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3299 OSSL_STORE_error and OSSL_STORE_close.
3300 The implementation uses backends called "loaders" to implement arbitrary
3301 URI schemes. There is one built in "loader" for the 'file' scheme.
3305 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3306 then adjusted to work on FreeBSD 8.4 as well.
3307 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3308 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3312 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3313 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3314 error code calls like this:
3316 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3318 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3319 that can be encoded in C. For the foreseeable future, this will only
3322 *Richard Levitte and Tim Hudson*
3324 * Removed BSD cryptodev engine.
3328 * Add a build target 'build_all_generated', to build all generated files
3329 and only that. This can be used to prepare everything that requires
3330 things like perl for a system that lacks perl and then move everything
3331 to that system and do the rest of the build there.
3335 * In the UI interface, make it possible to duplicate the user data. This
3336 can be used by engines that need to retain the data for a longer time
3337 than just the call where this user data is passed.
3341 * Ignore the '-named_curve auto' value for compatibility of applications
3344 *Tomáš Mráz <tmraz@fedoraproject.org>*
3346 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3347 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3348 alerts across multiple records (some of which could be empty). In practice
3349 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3350 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3351 support this at all. Supporting it adds significant complexity to the
3352 record layer, and its removal is unlikely to cause interoperability
3357 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3358 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3359 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3364 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3365 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3367 *Richard Levitte, Andy Polyakov*
3369 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3374 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3375 platform rather than 'mingw'.
3379 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3380 success if they are asked to add an object which already exists
3381 in the store. This change cascades to other functions which load
3382 certificates and CRLs.
3386 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3387 facilitate stack unwinding even from assembly subroutines.
3391 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3392 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3396 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3397 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3398 which is the minimum version we support.
3402 * Certificate time validation (X509_cmp_time) enforces stricter
3403 compliance with RFC 5280. Fractional seconds and timezone offsets
3404 are no longer allowed.
3408 * Add support for ARIA
3412 * s_client will now send the Server Name Indication (SNI) extension by
3413 default unless the new "-noservername" option is used. The server name is
3414 based on the host provided to the "-connect" option unless overridden by
3415 using "-servername".
3419 * Add support for SipHash
3423 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3424 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3425 prevent issues where no progress is being made and the peer continually
3426 sends unrecognised record types, using up resources processing them.
3430 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3431 using the algorithm defined in
3432 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3436 * Heartbeat support has been removed; the ABI is changed for now.
3438 *Richard Levitte, Rich Salz*
3440 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3444 * The RSA "null" method, which was partially supported to avoid patent
3445 issues, has been replaced to always returns NULL.
3452 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3454 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3455 used even when parsing explicit parameters, when loading a encoded key
3456 or calling `EC_GROUP_new_from_ecpkparameters()`/
3457 `EC_GROUP_new_from_ecparameters()`.
3458 This prevents bypass of security hardening and performance gains,
3459 especially for curves with specialized EC_METHODs.
3460 By default, if a key encoded with explicit parameters is loaded and later
3461 encoded, the output is still encoded with explicit parameters, even if
3462 internally a "named" EC_GROUP is used for computation.
3466 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3467 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3468 NULL. After this change, only the cofactor parameter can be NULL. It also
3469 does some minimal sanity checks on the passed order.
3474 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3475 An attack is simple, if the first CMS_recipientInfo is valid but the
3476 second CMS_recipientInfo is chosen ciphertext. If the second
3477 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3478 encryption key will be replaced by garbage, and the message cannot be
3479 decoded, but if the RSA decryption fails, the correct encryption key is
3480 used and the recipient will not notice the attack.
3481 As a work around for this potential attack the length of the decrypted
3482 key must be equal to the cipher default key length, in case the
3483 certificate is not given and all recipientInfo are tried out.
3484 The old behaviour can be re-enabled in the CMS code by setting the
3485 CMS_DEBUG_DECRYPT flag.
3490 * Use Windows installation paths in the mingw builds
3492 Mingw isn't a POSIX environment per se, which means that Windows
3493 paths should be used for installation.
3498 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3500 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3501 This changes the size when using the `genpkey` command when no size is given.
3502 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3503 generation commands to use 2048 bits by default.
3507 * Prevent over long nonces in ChaCha20-Poly1305.
3509 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3510 for every encryption operation. RFC 7539 specifies that the nonce value
3511 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3512 and front pads the nonce with 0 bytes if it is less than 12
3513 bytes. However it also incorrectly allows a nonce to be set of up to 16
3514 bytes. In this case only the last 12 bytes are significant and any
3515 additional leading bytes are ignored.
3517 It is a requirement of using this cipher that nonce values are
3518 unique. Messages encrypted using a reused nonce value are susceptible to
3519 serious confidentiality and integrity attacks. If an application changes
3520 the default nonce length to be longer than 12 bytes and then makes a
3521 change to the leading bytes of the nonce expecting the new value to be a
3522 new unique nonce then such an application could inadvertently encrypt
3523 messages with a reused nonce.
3525 Additionally the ignored bytes in a long nonce are not covered by the
3526 integrity guarantee of this cipher. Any application that relies on the
3527 integrity of these ignored leading bytes of a long nonce may be further
3528 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3529 is safe because no such use sets such a long nonce value. However user
3530 applications that use this cipher directly and set a non-default nonce
3531 length to be longer than 12 bytes may be vulnerable.
3533 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3539 * Added SCA hardening for modular field inversion in EC_GROUP through
3540 a new dedicated field_inv() pointer in EC_METHOD.
3541 This also addresses a leakage affecting conversions from projective
3542 to affine coordinates.
3544 *Billy Bob Brumley, Nicola Tuveri*
3546 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3547 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3551 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3555 * Remove the 'dist' target and add a tarball building script. The
3556 'dist' target has fallen out of use, and it shouldn't be
3557 necessary to configure just to create a source distribution.
3561 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3563 * Timing vulnerability in DSA signature generation
3565 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3566 timing side channel attack. An attacker could use variations in the signing
3567 algorithm to recover the private key.
3569 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3574 * Timing vulnerability in ECDSA signature generation
3576 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3577 timing side channel attack. An attacker could use variations in the signing
3578 algorithm to recover the private key.
3580 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3585 * Add coordinate blinding for EC_POINT and implement projective
3586 coordinate blinding for generic prime curves as a countermeasure to
3587 chosen point SCA attacks.
3589 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3591 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3593 * Client DoS due to large DH parameter
3595 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3596 malicious server can send a very large prime value to the client. This will
3597 cause the client to spend an unreasonably long period of time generating a
3598 key for this prime resulting in a hang until the client has finished. This
3599 could be exploited in a Denial Of Service attack.
3601 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3606 * Cache timing vulnerability in RSA Key Generation
3608 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3609 a cache timing side channel attack. An attacker with sufficient access to
3610 mount cache timing attacks during the RSA key generation process could
3611 recover the private key.
3613 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3614 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3619 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3620 parameter is no longer accepted, as it leads to a corrupt table. NULL
3621 pem_str is reserved for alias entries only.
3625 * Revert blinding in ECDSA sign and instead make problematic addition
3626 length-invariant. Switch even to fixed-length Montgomery multiplication.
3630 * Change generating and checking of primes so that the error rate of not
3631 being prime depends on the intended use based on the size of the input.
3632 For larger primes this will result in more rounds of Miller-Rabin.
3633 The maximal error rate for primes with more than 1080 bits is lowered
3636 *Kurt Roeckx, Annie Yousar*
3638 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3642 * Add blinding to ECDSA and DSA signatures to protect against side channel
3643 attacks discovered by Keegan Ryan (NCC Group).
3647 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3648 now allow empty (zero character) pass phrases.
3652 * Certificate time validation (X509_cmp_time) enforces stricter
3653 compliance with RFC 5280. Fractional seconds and timezone offsets
3654 are no longer allowed.
3658 * Fixed a text canonicalisation bug in CMS
3660 Where a CMS detached signature is used with text content the text goes
3661 through a canonicalisation process first prior to signing or verifying a
3662 signature. This process strips trailing space at the end of lines, converts
3663 line terminators to CRLF and removes additional trailing line terminators
3664 at the end of a file. A bug in the canonicalisation process meant that
3665 some characters, such as form-feed, were incorrectly treated as whitespace
3666 and removed. This is contrary to the specification (RFC5485). This fix
3667 could mean that detached text data signed with an earlier version of
3668 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3669 signed with a fixed OpenSSL may fail to verify with an earlier version of
3670 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3671 and use the "-binary" flag (for the "cms" command line application) or set
3672 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3676 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3678 * Constructed ASN.1 types with a recursive definition could exceed the stack
3680 Constructed ASN.1 types with a recursive definition (such as can be found
3681 in PKCS7) could eventually exceed the stack given malicious input with
3682 excessive recursion. This could result in a Denial Of Service attack. There
3683 are no such structures used within SSL/TLS that come from untrusted sources
3684 so this is considered safe.
3686 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3692 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3694 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3695 effectively reduced to only comparing the least significant bit of each
3696 byte. This allows an attacker to forge messages that would be considered as
3697 authenticated in an amount of tries lower than that guaranteed by the
3698 security claims of the scheme. The module can only be compiled by the
3699 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3701 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3707 * Add a build target 'build_all_generated', to build all generated files
3708 and only that. This can be used to prepare everything that requires
3709 things like perl for a system that lacks perl and then move everything
3710 to that system and do the rest of the build there.
3714 * Backport SSL_OP_NO_RENGOTIATION
3716 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3717 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3718 changes this is no longer possible in 1.1.0. Therefore, the new
3719 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3720 1.1.0 to provide equivalent functionality.
3722 Note that if an application built against 1.1.0h headers (or above) is run
3723 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3724 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3728 * Removed the OS390-Unix config target. It relied on a script that doesn't
3733 * rsaz_1024_mul_avx2 overflow bug on x86_64
3735 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3736 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3737 Analysis suggests that attacks against RSA and DSA as a result of this
3738 defect would be very difficult to perform and are not believed likely.
3739 Attacks against DH1024 are considered just feasible, because most of the
3740 work necessary to deduce information about a private key may be performed
3741 offline. The amount of resources required for such an attack would be
3742 significant. However, for an attack on TLS to be meaningful, the server
3743 would have to share the DH1024 private key among multiple clients, which is
3744 no longer an option since CVE-2016-0701.
3746 This only affects processors that support the AVX2 but not ADX extensions
3747 like Intel Haswell (4th generation).
3749 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3750 was originally found via the OSS-Fuzz project.
3755 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3757 * bn_sqrx8x_internal carry bug on x86_64
3759 There is a carry propagating bug in the x86_64 Montgomery squaring
3760 procedure. No EC algorithms are affected. Analysis suggests that attacks
3761 against RSA and DSA as a result of this defect would be very difficult to
3762 perform and are not believed likely. Attacks against DH are considered just
3763 feasible (although very difficult) because most of the work necessary to
3764 deduce information about a private key may be performed offline. The amount
3765 of resources required for such an attack would be very significant and
3766 likely only accessible to a limited number of attackers. An attacker would
3767 additionally need online access to an unpatched system using the target
3768 private key in a scenario with persistent DH parameters and a private
3769 key that is shared between multiple clients.
3771 This only affects processors that support the BMI1, BMI2 and ADX extensions
3772 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3774 This issue was reported to OpenSSL by the OSS-Fuzz project.
3779 * Malformed X.509 IPAddressFamily could cause OOB read
3781 If an X.509 certificate has a malformed IPAddressFamily extension,
3782 OpenSSL could do a one-byte buffer overread. The most likely result
3783 would be an erroneous display of the certificate in text format.
3785 This issue was reported to OpenSSL by the OSS-Fuzz project.
3790 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3792 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3793 platform rather than 'mingw'.
3797 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3798 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3799 which is the minimum version we support.
3803 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3805 * Encrypt-Then-Mac renegotiation crash
3807 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3808 negotiated where it was not in the original handshake (or vice-versa) then
3809 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3810 and servers are affected.
3812 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3817 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3819 * Truncated packet could crash via OOB read
3821 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3822 cipher is being used, then a truncated packet can cause that host to
3823 perform an out-of-bounds read, usually resulting in a crash.
3825 This issue was reported to OpenSSL by Robert Święcki of Google.
3830 * Bad (EC)DHE parameters cause a client crash
3832 If a malicious server supplies bad parameters for a DHE or ECDHE key
3833 exchange then this can result in the client attempting to dereference a
3834 NULL pointer leading to a client crash. This could be exploited in a Denial
3837 This issue was reported to OpenSSL by Guido Vranken.
3842 * BN_mod_exp may produce incorrect results on x86_64
3844 There is a carry propagating bug in the x86_64 Montgomery squaring
3845 procedure. No EC algorithms are affected. Analysis suggests that attacks
3846 against RSA and DSA as a result of this defect would be very difficult to
3847 perform and are not believed likely. Attacks against DH are considered just
3848 feasible (although very difficult) because most of the work necessary to
3849 deduce information about a private key may be performed offline. The amount
3850 of resources required for such an attack would be very significant and
3851 likely only accessible to a limited number of attackers. An attacker would
3852 additionally need online access to an unpatched system using the target
3853 private key in a scenario with persistent DH parameters and a private
3854 key that is shared between multiple clients. For example this can occur by
3855 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3856 similar to CVE-2015-3193 but must be treated as a separate problem.
3858 This issue was reported to OpenSSL by the OSS-Fuzz project.
3863 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3865 * ChaCha20/Poly1305 heap-buffer-overflow
3867 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3868 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3869 crash. This issue is not considered to be exploitable beyond a DoS.
3871 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3876 * CMS Null dereference
3878 Applications parsing invalid CMS structures can crash with a NULL pointer
3879 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3880 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3881 structure callback if an attempt is made to free certain invalid encodings.
3882 Only CHOICE structures using a callback which do not handle NULL value are
3885 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3890 * Montgomery multiplication may produce incorrect results
3892 There is a carry propagating bug in the Broadwell-specific Montgomery
3893 multiplication procedure that handles input lengths divisible by, but
3894 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3895 and DH private keys are impossible. This is because the subroutine in
3896 question is not used in operations with the private key itself and an input
3897 of the attacker's direct choice. Otherwise the bug can manifest itself as
3898 transient authentication and key negotiation failures or reproducible
3899 erroneous outcome of public-key operations with specially crafted input.
3900 Among EC algorithms only Brainpool P-512 curves are affected and one
3901 presumably can attack ECDH key negotiation. Impact was not analyzed in
3902 detail, because pre-requisites for attack are considered unlikely. Namely
3903 multiple clients have to choose the curve in question and the server has to
3904 share the private key among them, neither of which is default behaviour.
3905 Even then only clients that chose the curve will be affected.
3907 This issue was publicly reported as transient failures and was not
3908 initially recognized as a security issue. Thanks to Richard Morgan for
3909 providing reproducible case.
3914 * Removed automatic addition of RPATH in shared libraries and executables,
3915 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3919 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3921 * Fix Use After Free for large message sizes
3923 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3924 message larger than approx 16k is received then the underlying buffer to
3925 store the incoming message is reallocated and moved. Unfortunately a
3926 dangling pointer to the old location is left which results in an attempt to
3927 write to the previously freed location. This is likely to result in a
3928 crash, however it could potentially lead to execution of arbitrary code.
3930 This issue only affects OpenSSL 1.1.0a.
3932 This issue was reported to OpenSSL by Robert Święcki.
3937 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3939 * OCSP Status Request extension unbounded memory growth
3941 A malicious client can send an excessively large OCSP Status Request
3942 extension. If that client continually requests renegotiation, sending a
3943 large OCSP Status Request extension each time, then there will be unbounded
3944 memory growth on the server. This will eventually lead to a Denial Of
3945 Service attack through memory exhaustion. Servers with a default
3946 configuration are vulnerable even if they do not support OCSP. Builds using
3947 the "no-ocsp" build time option are not affected.
3949 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3954 * SSL_peek() hang on empty record
3956 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3957 sends an empty record. This could be exploited by a malicious peer in a
3958 Denial Of Service attack.
3960 This issue was reported to OpenSSL by Alex Gaynor.
3965 * Excessive allocation of memory in tls_get_message_header() and
3966 dtls1_preprocess_fragment()
3968 A (D)TLS message includes 3 bytes for its length in the header for the
3969 message. This would allow for messages up to 16Mb in length. Messages of
3970 this length are excessive and OpenSSL includes a check to ensure that a
3971 peer is sending reasonably sized messages in order to avoid too much memory
3972 being consumed to service a connection. A flaw in the logic of version
3973 1.1.0 means that memory for the message is allocated too early, prior to
3974 the excessive message length check. Due to way memory is allocated in
3975 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3976 to service a connection. This could lead to a Denial of Service through
3977 memory exhaustion. However, the excessive message length check still takes
3978 place, and this would cause the connection to immediately fail. Assuming
3979 that the application calls SSL_free() on the failed connection in a timely
3980 manner then the 21Mb of allocated memory will then be immediately freed
3981 again. Therefore, the excessive memory allocation will be transitory in
3982 nature. This then means that there is only a security impact if:
3984 1) The application does not call SSL_free() in a timely manner in the event
3985 that the connection fails
3987 2) The application is working in a constrained environment where there is
3988 very little free memory
3990 3) The attacker initiates multiple connection attempts such that there are
3991 multiple connections in a state where memory has been allocated for the
3992 connection; SSL_free() has not yet been called; and there is insufficient
3993 memory to service the multiple requests.
3995 Except in the instance of (1) above any Denial Of Service is likely to be
3996 transitory because as soon as the connection fails the memory is
3997 subsequently freed again in the SSL_free() call. However there is an
3998 increased risk during this period of application crashes due to the lack of
3999 memory - which would then mean a more serious Denial of Service.
4001 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4002 (CVE-2016-6307 and CVE-2016-6308)
4006 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4007 had to be removed. Primary reason is that vendor assembler can't
4008 assemble our modules with -KPIC flag. As result it, assembly
4009 support, was not even available as option. But its lack means
4010 lack of side-channel resistant code, which is incompatible with
4011 security by todays standards. Fortunately gcc is readily available
4012 prepackaged option, which we firmly point at...
4016 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4018 * Windows command-line tool supports UTF-8 opt-in option for arguments
4019 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4020 (to any value) allows Windows user to access PKCS#12 file generated
4021 with Windows CryptoAPI and protected with non-ASCII password, as well
4022 as files generated under UTF-8 locale on Linux also protected with
4027 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4028 have been disabled by default and removed from DEFAULT, just like RC4.
4029 See the RC4 item below to re-enable both.
4033 * The method for finding the storage location for the Windows RAND seed file
4034 has changed. First we check %RANDFILE%. If that is not set then we check
4035 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4036 all else fails we fall back to C:\.
4040 * The EVP_EncryptUpdate() function has had its return type changed from void
4041 to int. A return of 0 indicates and error while a return of 1 indicates
4046 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4047 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4048 off the constant time implementation for RSA, DSA and DH have been made
4049 no-ops and deprecated.
4053 * Windows RAND implementation was simplified to only get entropy by
4054 calling CryptGenRandom(). Various other RAND-related tickets
4057 *Joseph Wylie Yandle, Rich Salz*
4059 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4060 and `OPENSSL_LH_`, respectively. The old names are available
4061 with API compatibility. They new names are now completely documented.
4065 * Unify TYPE_up_ref(obj) methods signature.
4066 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4067 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4068 int (instead of void) like all others TYPE_up_ref() methods.
4069 So now these methods also check the return value of CRYPTO_atomic_add(),
4070 and the validity of object reference counter.
4072 *fdasilvayy@gmail.com*
4074 * With Windows Visual Studio builds, the .pdb files are installed
4075 alongside the installed libraries and executables. For a static
4076 library installation, ossl_static.pdb is the associate compiler
4077 generated .pdb file to be used when linking programs.
4081 * Remove openssl.spec. Packaging files belong with the packagers.
4085 * Automatic Darwin/OSX configuration has had a refresh, it will now
4086 recognise x86_64 architectures automatically. You can still decide
4087 to build for a different bitness with the environment variable
4088 KERNEL_BITS (can be 32 or 64), for example:
4090 KERNEL_BITS=32 ./config
4094 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4095 256 bit AES and HMAC with SHA256.
4099 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4103 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4107 * To enable users to have their own config files and build file templates,
4108 Configure looks in the directory indicated by the environment variable
4109 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4110 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4111 name and is used as is.
4115 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4116 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4117 X509_CERT_FILE_CTX was removed.
4121 * "shared" builds are now the default. To create only static libraries use
4122 the "no-shared" Configure option.
4126 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4127 All of these option have not worked for some while and are fundamental
4132 * Make various cleanup routines no-ops and mark them as deprecated. Most
4133 global cleanup functions are no longer required because they are handled
4134 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4135 Explicitly de-initing can cause problems (e.g. where a library that uses
4136 OpenSSL de-inits, but an application is still using it). The affected
4137 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4138 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4139 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4140 COMP_zlib_cleanup().
4144 * --strict-warnings no longer enables runtime debugging options
4145 such as REF_DEBUG. Instead, debug options are automatically
4146 enabled with '--debug' builds.
4148 *Andy Polyakov, Emilia Käsper*
4150 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4151 have been moved out of the public header files. New functions for managing
4152 these have been added.
4156 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4157 objects have been moved out of the public header files. New
4158 functions for managing these have been added.
4162 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4163 have been moved out of the public header files. New functions for managing
4164 these have been added.
4168 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4169 moved out of the public header files. New functions for managing these
4174 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4178 * Removed the mk1mf build scripts.
4182 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4183 it is always safe to #include a header now.
4187 * Removed the aged BC-32 config and all its supporting scripts
4191 * Removed support for Ultrix, Netware, and OS/2.
4195 * Add support for HKDF.
4197 *Alessandro Ghedini*
4199 * Add support for blake2b and blake2s
4203 * Added support for "pipelining". Ciphers that have the
4204 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4205 encryptions/decryptions simultaneously. There are currently no built-in
4206 ciphers with this property but the expectation is that engines will be able
4207 to offer it to significantly improve throughput. Support has been extended
4208 into libssl so that multiple records for a single connection can be
4209 processed in one go (for >=TLS 1.1).
4213 * Added the AFALG engine. This is an async capable engine which is able to
4214 offload work to the Linux kernel. In this initial version it only supports
4215 AES128-CBC. The kernel must be version 4.1.0 or greater.
4219 * OpenSSL now uses a new threading API. It is no longer necessary to
4220 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4221 are two supported threading models: pthreads and windows threads. It is
4222 also possible to configure OpenSSL at compile time for "no-threads". The
4223 old threading API should no longer be used. The functions have been
4224 replaced with "no-op" compatibility macros.
4226 *Alessandro Ghedini, Matt Caswell*
4228 * Modify behavior of ALPN to invoke callback after SNI/servername
4229 callback, such that updates to the SSL_CTX affect ALPN.
4233 * Add SSL_CIPHER queries for authentication and key-exchange.
4237 * Changes to the DEFAULT cipherlist:
4238 - Prefer (EC)DHE handshakes over plain RSA.
4239 - Prefer AEAD ciphers over legacy ciphers.
4240 - Prefer ECDSA over RSA when both certificates are available.
4241 - Prefer TLSv1.2 ciphers/PRF.
4242 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4247 * Change the ECC default curve list to be this, in order: x25519,
4248 secp256r1, secp521r1, secp384r1.
4252 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4253 disabled by default. They can be re-enabled using the
4254 enable-weak-ssl-ciphers option to Configure.
4258 * If the server has ALPN configured, but supports no protocols that the
4259 client advertises, send a fatal "no_application_protocol" alert.
4260 This behaviour is SHALL in RFC 7301, though it isn't universally
4261 implemented by other servers.
4265 * Add X25519 support.
4266 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4267 for public and private key encoding using the format documented in
4268 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4269 key generation and key derivation.
4271 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4276 * Deprecate SRP_VBASE_get_by_user.
4277 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4278 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4279 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4280 seed, even if the seed is configured.
4282 Users should use SRP_VBASE_get1_by_user instead. Note that in
4283 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4284 also that even though configuring the SRP seed attempts to hide
4285 invalid usernames by continuing the handshake with fake
4286 credentials, this behaviour is not constant time and no strong
4287 guarantees are made that the handshake is indistinguishable from
4288 that of a valid user.
4292 * Configuration change; it's now possible to build dynamic engines
4293 without having to build shared libraries and vice versa. This
4294 only applies to the engines in `engines/`, those in `crypto/engine/`
4295 will always be built into libcrypto (i.e. "static").
4297 Building dynamic engines is enabled by default; to disable, use
4298 the configuration option "disable-dynamic-engine".
4300 The only requirements for building dynamic engines are the
4301 presence of the DSO module and building with position independent
4302 code, so they will also automatically be disabled if configuring
4303 with "disable-dso" or "disable-pic".
4305 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4306 are also taken away from openssl/opensslconf.h, as they are
4311 * Configuration change; if there is a known flag to compile
4312 position independent code, it will always be applied on the
4313 libcrypto and libssl object files, and never on the application
4314 object files. This means other libraries that use routines from
4315 libcrypto / libssl can be made into shared libraries regardless
4316 of how OpenSSL was configured.
4318 If this isn't desirable, the configuration options "disable-pic"
4319 or "no-pic" can be used to disable the use of PIC. This will
4320 also disable building shared libraries and dynamic engines.
4324 * Removed JPAKE code. It was experimental and has no wide use.
4328 * The INSTALL_PREFIX Makefile variable has been renamed to
4329 DESTDIR. That makes for less confusion on what this variable
4330 is for. Also, the configuration option --install_prefix is
4335 * Heartbeat for TLS has been removed and is disabled by default
4336 for DTLS; configure with enable-heartbeats. Code that uses the
4337 old #define's might need to be updated.
4339 *Emilia Käsper, Rich Salz*
4341 * Rename REF_CHECK to REF_DEBUG.
4345 * New "unified" build system
4347 The "unified" build system is aimed to be a common system for all
4348 platforms we support. With it comes new support for VMS.
4350 This system builds supports building in a different directory tree
4351 than the source tree. It produces one Makefile (for unix family
4352 or lookalikes), or one descrip.mms (for VMS).
4354 The source of information to make the Makefile / descrip.mms is
4355 small files called 'build.info', holding the necessary
4356 information for each directory with source to compile, and a
4357 template in Configurations, like unix-Makefile.tmpl or
4360 With this change, the library names were also renamed on Windows
4361 and on VMS. They now have names that are closer to the standard
4362 on Unix, and include the major version number, and in certain
4363 cases, the architecture they are built for. See "Notes on shared
4364 libraries" in INSTALL.
4366 We rely heavily on the perl module Text::Template.
4370 * Added support for auto-initialisation and de-initialisation of the library.
4371 OpenSSL no longer requires explicit init or deinit routines to be called,
4372 except in certain circumstances. See the OPENSSL_init_crypto() and
4373 OPENSSL_init_ssl() man pages for further information.
4377 * The arguments to the DTLSv1_listen function have changed. Specifically the
4378 "peer" argument is now expected to be a BIO_ADDR object.
4380 * Rewrite of BIO networking library. The BIO library lacked consistent
4381 support of IPv6, and adding it required some more extensive
4382 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4383 which hold all types of addresses and chains of address information.
4384 It also introduces a new API, with functions like BIO_socket,
4385 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4386 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4387 have been adapted accordingly.
4391 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4396 * CRIME protection: disable compression by default, even if OpenSSL is
4397 compiled with zlib enabled. Applications can still enable compression
4398 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4399 using the SSL_CONF library to configure compression.
4403 * The signature of the session callback configured with
4404 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4405 was explicitly marked as `const unsigned char*` instead of
4410 * Always DPURIFY. Remove the use of uninitialized memory in the
4411 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4415 * Removed many obsolete configuration items, including
4416 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4417 MD2_CHAR, MD2_INT, MD2_LONG
4419 IDEA_SHORT, IDEA_LONG
4420 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4422 *Rich Salz, with advice from Andy Polyakov*
4424 * Many BN internals have been moved to an internal header file.
4426 *Rich Salz with help from Andy Polyakov*
4428 * Configuration and writing out the results from it has changed.
4429 Files such as Makefile include/openssl/opensslconf.h and are now
4430 produced through general templates, such as Makefile.in and
4431 crypto/opensslconf.h.in and some help from the perl module
4434 Also, the center of configuration information is no longer
4435 Makefile. Instead, Configure produces a perl module in
4436 configdata.pm which holds most of the config data (in the hash
4437 table %config), the target data that comes from the target
4438 configuration in one of the `Configurations/*.conf` files (in
4443 * To clarify their intended purposes, the Configure options
4444 --prefix and --openssldir change their semantics, and become more
4445 straightforward and less interdependent.
4447 --prefix shall be used exclusively to give the location INSTALLTOP
4448 where programs, scripts, libraries, include files and manuals are
4449 going to be installed. The default is now /usr/local.
4451 --openssldir shall be used exclusively to give the default
4452 location OPENSSLDIR where certificates, private keys, CRLs are
4453 managed. This is also where the default openssl.cnf gets
4455 If the directory given with this option is a relative path, the
4456 values of both the --prefix value and the --openssldir value will
4457 be combined to become OPENSSLDIR.
4458 The default for --openssldir is INSTALLTOP/ssl.
4460 Anyone who uses --openssldir to specify where OpenSSL is to be
4461 installed MUST change to use --prefix instead.
4465 * The GOST engine was out of date and therefore it has been removed. An up
4466 to date GOST engine is now being maintained in an external repository.
4467 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4468 support for GOST ciphersuites (these are only activated if a GOST engine
4473 * EGD is no longer supported by default; use enable-egd when
4476 *Ben Kaduk and Rich Salz*
4478 * The distribution now has Makefile.in files, which are used to
4479 create Makefile's when Configure is run. *Configure must be run
4480 before trying to build now.*
4484 * The return value for SSL_CIPHER_description() for error conditions
4489 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4491 Obtaining and performing DNSSEC validation of TLSA records is
4492 the application's responsibility. The application provides
4493 the TLSA records of its choice to OpenSSL, and these are then
4494 used to authenticate the peer.
4496 The TLSA records need not even come from DNS. They can, for
4497 example, be used to implement local end-entity certificate or
4498 trust-anchor "pinning", where the "pin" data takes the form
4499 of TLSA records, which can augment or replace verification
4500 based on the usual WebPKI public certification authorities.
4504 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4505 continues to support deprecated interfaces in default builds.
4506 However, applications are strongly advised to compile their
4507 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4508 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4509 or the 1.1.0 releases.
4511 In environments in which all applications have been ported to
4512 not use any deprecated interfaces OpenSSL's Configure script
4513 should be used with the --api=1.1.0 option to entirely remove
4514 support for the deprecated features from the library and
4515 unconditionally disable them in the installed headers.
4516 Essentially the same effect can be achieved with the "no-deprecated"
4517 argument to Configure, except that this will always restrict
4518 the build to just the latest API, rather than a fixed API
4521 As applications are ported to future revisions of the API,
4522 they should update their compile-time OPENSSL_API_COMPAT define
4523 accordingly, but in most cases should be able to continue to
4524 compile with later releases.
4526 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4527 0x10000000L and 0x00908000L, respectively. However those
4528 versions did not support the OPENSSL_API_COMPAT feature, and
4529 so applications are not typically tested for explicit support
4530 of just the undeprecated features of either release.
4534 * Add support for setting the minimum and maximum supported protocol.
4535 It can bet set via the SSL_set_min_proto_version() and
4536 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4537 MaxProtocol. It's recommended to use the new APIs to disable
4538 protocols instead of disabling individual protocols using
4539 SSL_set_options() or SSL_CONF's Protocol. This change also
4540 removes support for disabling TLS 1.2 in the OpenSSL TLS
4541 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4545 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4549 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4550 and integrates ECDSA and ECDH functionality into EC. Implementations can
4551 now redirect key generation and no longer need to convert to or from
4554 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4555 include the ec.h header file instead.
4559 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4560 ciphers who are no longer supported and drops support the ephemeral RSA key
4561 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4565 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4566 opaque. For HMAC_CTX, the following constructors and destructors
4569 HMAC_CTX *HMAC_CTX_new(void);
4570 void HMAC_CTX_free(HMAC_CTX *ctx);
4572 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4573 destroy such methods has been added. See EVP_MD_meth_new(3) and
4574 EVP_CIPHER_meth_new(3) for documentation.
4577 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4578 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4579 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4580 an already created structure.
4581 2) For consistency with the majority of our object creators and
4582 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4583 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4584 for deprecated builds.
4588 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4589 cryptographic operations to be performed asynchronously as long as an
4590 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4591 further details. Libssl has also had this capability integrated with the
4592 introduction of the new mode SSL_MODE_ASYNC and associated error
4593 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4594 pages. This work was developed in partnership with Intel Corp.
4598 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4599 always enabled now. If you want to disable the support you should
4600 exclude it using the list of supported ciphers. This also means that the
4601 "-no_ecdhe" option has been removed from s_server.
4605 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4606 SSL_{CTX_}set1_curves() which can set a list.
4610 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4611 curve you want to support using SSL_{CTX_}set1_curves().
4615 * State machine rewrite. The state machine code has been significantly
4616 refactored in order to remove much duplication of code and solve issues
4617 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4618 further details). This change does have some associated API changes.
4619 Notably the SSL_state() function has been removed and replaced by
4620 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4621 SSL_set_state() has been removed altogether. The previous handshake states
4622 defined in ssl.h and ssl3.h have also been removed.
4626 * All instances of the string "ssleay" in the public API were replaced
4627 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4628 Some error codes related to internal RSA_eay API's were renamed.
4632 * The demo files in crypto/threads were moved to demo/threads.
4636 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4639 *Matt Caswell, Rich Salz*
4641 * New ASN.1 embed macro.
4643 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4644 structure is not allocated: it is part of the parent. That is instead of
4652 This reduces memory fragmentation and make it impossible to accidentally
4653 set a mandatory field to NULL.
4655 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4656 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4657 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4662 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4666 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4667 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4668 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4669 DES and RC4 ciphersuites.
4673 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4674 This changes the decoding behaviour for some invalid messages,
4675 though the change is mostly in the more lenient direction, and
4676 legacy behaviour is preserved as much as possible.
4680 * Fix no-stdio build.
4681 *David Woodhouse <David.Woodhouse@intel.com> and also*
4682 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4684 * New testing framework
4685 The testing framework has been largely rewritten and is now using
4686 perl and the perl modules Test::Harness and an extended variant of
4687 Test::More called OpenSSL::Test to do its work. All test scripts in
4688 test/ have been rewritten into test recipes, and all direct calls to
4689 executables in test/Makefile have become individual recipes using the
4690 simplified testing OpenSSL::Test::Simple.
4692 For documentation on our testing modules, do:
4694 perldoc test/testlib/OpenSSL/Test/Simple.pm
4695 perldoc test/testlib/OpenSSL/Test.pm
4699 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4700 are used; the latter aborts on memory leaks (usually checked on exit).
4701 Some undocumented "set malloc, etc., hooks" functions were removed
4702 and others were changed. All are now documented.
4706 * In DSA_generate_parameters_ex, if the provided seed is too short,
4709 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4711 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4712 from RFC4279, RFC4785, RFC5487, RFC5489.
4714 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4715 original RSA_PSK patch.
4719 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4720 era flag was never set throughout the codebase (only read). Also removed
4721 SSL3_FLAGS_POP_BUFFER which was only used if
4722 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4726 * Changed the default name options in the "ca", "crl", "req" and "x509"
4727 to be "oneline" instead of "compat".
4731 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4732 not aware of clients that still exhibit this bug, and the workaround
4733 hasn't been working properly for a while.
4737 * The return type of BIO_number_read() and BIO_number_written() as well as
4738 the corresponding num_read and num_write members in the BIO structure has
4739 changed from unsigned long to uint64_t. On platforms where an unsigned
4740 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4745 * Given the pervasive nature of TLS extensions it is inadvisable to run
4746 OpenSSL without support for them. It also means that maintaining
4747 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4748 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4752 * Removed support for the two export grade static DH ciphersuites
4753 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4754 were newly added (along with a number of other static DH ciphersuites) to
4755 1.0.2. However the two export ones have *never* worked since they were
4756 introduced. It seems strange in any case to be adding new export
4757 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4761 * Version negotiation has been rewritten. In particular SSLv23_method(),
4762 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4763 and turned into macros which simply call the new preferred function names
4764 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4765 should use the new names instead. Also as part of this change the ssl23.h
4766 header file has been removed.
4770 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4771 code and the associated standard is no longer considered fit-for-purpose.
4775 * RT2547 was closed. When generating a private key, try to make the
4776 output file readable only by the owner. This behavior change might
4777 be noticeable when interacting with other software.
4779 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4784 * Added HTTP GET support to the ocsp command.
4788 * Changed default digest for the dgst and enc commands from MD5 to
4793 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4797 * Added support for TLS extended master secret from
4798 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4799 initial patch which was a great help during development.
4803 * All libssl internal structures have been removed from the public header
4804 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4805 now redundant). Users should not attempt to access internal structures
4806 directly. Instead they should use the provided API functions.
4810 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4811 Access to deprecated functions can be re-enabled by running config with
4812 "enable-deprecated". In addition applications wishing to use deprecated
4813 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4814 will, by default, disable some transitive includes that previously existed
4815 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4819 * Added support for OCB mode. OpenSSL has been granted a patent license
4820 compatible with the OpenSSL license for use of OCB. Details are available
4821 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4822 for OCB can be removed by calling config with no-ocb.
4826 * SSLv2 support has been removed. It still supports receiving an SSLv2
4827 compatible client hello.
4831 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4832 done while fixing the error code for the key-too-small case.
4834 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4836 * CA.sh has been removed; use CA.pl instead.
4840 * Removed old DES API.
4844 * Remove various unsupported platforms:
4850 Sinix/ReliantUNIX RM400
4855 16-bit platforms such as WIN16
4859 * Clean up OPENSSL_NO_xxx #define's
4860 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4861 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4862 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4863 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4864 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4865 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4866 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4867 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4868 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4869 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4873 * Cleaned up dead code
4874 Remove all but one '#ifdef undef' which is to be looked at.
4878 * Clean up calling of xxx_free routines.
4879 Just like free(), fix most of the xxx_free routines to accept
4880 NULL. Remove the non-null checks from callers. Save much code.
4884 * Add secure heap for storage of private keys (when possible).
4885 Add BIO_s_secmem(), CBIGNUM, etc.
4886 Contributed by Akamai Technologies under our Corporate CLA.
4890 * Experimental support for a new, fast, unbiased prime candidate generator,
4891 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4893 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4895 * New output format NSS in the sess_id command line tool. This allows
4896 exporting the session id and the master key in NSS keylog format.
4898 *Martin Kaiser <martin@kaiser.cx>*
4900 * Harmonize version and its documentation. -f flag is used to display
4903 *mancha <mancha1@zoho.com>*
4905 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4906 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4908 *mancha <mancha1@zoho.com>*
4910 * Fix some double frees. These are not thought to be exploitable.
4912 *mancha <mancha1@zoho.com>*
4914 * A missing bounds check in the handling of the TLS heartbeat extension
4915 can be used to reveal up to 64k of memory to a connected client or
4918 Thanks for Neel Mehta of Google Security for discovering this bug and to
4919 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4920 preparing the fix ([CVE-2014-0160])
4922 *Adam Langley, Bodo Moeller*
4924 * Fix for the attack described in the paper "Recovering OpenSSL
4925 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4926 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4927 <http://eprint.iacr.org/2014/140>
4929 Thanks to Yuval Yarom and Naomi Benger for discovering this
4930 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4932 *Yuval Yarom and Naomi Benger*
4934 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4935 this fixes a limitation in previous versions of OpenSSL.
4939 * Experimental encrypt-then-mac support.
4941 Experimental support for encrypt then mac from
4942 draft-gutmann-tls-encrypt-then-mac-02.txt
4944 To enable it set the appropriate extension number (0x42 for the test
4945 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4947 For non-compliant peers (i.e. just about everything) this should have no
4950 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4954 * Add EVP support for key wrapping algorithms, to avoid problems with
4955 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4956 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4957 algorithms and include tests cases.
4961 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4966 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4967 MGF1 digest and OAEP label.
4971 * Make openssl verify return errors.
4973 *Chris Palmer <palmer@google.com> and Ben Laurie*
4975 * New function ASN1_TIME_diff to calculate the difference between two
4976 ASN1_TIME structures or one structure and the current time.
4980 * Update fips_test_suite to support multiple command line options. New
4981 test to induce all self test errors in sequence and check expected
4986 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4987 sign or verify all in one operation.
4991 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4992 test programs and fips_test_suite. Includes functionality to parse
4993 the minimal script output of fipsalgest.pl directly.
4997 * Add authorisation parameter to FIPS_module_mode_set().
5001 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5005 * Use separate DRBG fields for internal and external flags. New function
5006 FIPS_drbg_health_check() to perform on demand health checking. Add
5007 generation tests to fips_test_suite with reduced health check interval to
5008 demonstrate periodic health checking. Add "nodh" option to
5009 fips_test_suite to skip very slow DH test.
5013 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5018 * More extensive health check for DRBG checking many more failure modes.
5019 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5020 combination: call this in fips_test_suite.
5024 * Add support for canonical generation of DSA parameter 'g'. See
5027 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5028 POST to handle HMAC cases.
5032 * Add functions FIPS_module_version() and FIPS_module_version_text()
5033 to return numerical and string versions of the FIPS module number.
5037 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5038 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5039 outside the validated module in the FIPS capable OpenSSL.
5043 * Minor change to DRBG entropy callback semantics. In some cases
5044 there is no multiple of the block length between min_len and
5045 max_len. Allow the callback to return more than max_len bytes
5046 of entropy but discard any extra: it is the callback's responsibility
5047 to ensure that the extra data discarded does not impact the
5048 requested amount of entropy.
5052 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5053 information in FIPS186-3, SP800-57 and SP800-131A.
5057 * CCM support via EVP. Interface is very similar to GCM case except we
5058 must supply all data in one chunk (i.e. no update, final) and the
5059 message length must be supplied if AAD is used. Add algorithm test
5064 * Initial version of POST overhaul. Add POST callback to allow the status
5065 of POST to be monitored and/or failures induced. Modify fips_test_suite
5066 to use callback. Always run all selftests even if one fails.
5070 * XTS support including algorithm test driver in the fips_gcmtest program.
5071 Note: this does increase the maximum key length from 32 to 64 bytes but
5072 there should be no binary compatibility issues as existing applications
5073 will never use XTS mode.
5077 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5078 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5079 performs algorithm blocking for unapproved PRNG types. Also do not
5080 set PRNG type in FIPS_mode_set(): leave this to the application.
5081 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5082 the standard OpenSSL PRNG: set additional data to a date time vector.
5086 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5087 This shouldn't present any incompatibility problems because applications
5088 shouldn't be using these directly and any that are will need to rethink
5089 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5093 * Extensive self tests and health checking required by SP800-90 DRBG.
5094 Remove strength parameter from FIPS_drbg_instantiate and always
5095 instantiate at maximum supported strength.
5099 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5103 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5107 * New function DH_compute_key_padded() to compute a DH key and pad with
5108 leading zeroes if needed: this complies with SP800-56A et al.
5112 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5113 anything, incomplete, subject to change and largely untested at present.
5117 * Modify fipscanisteronly build option to only build the necessary object
5118 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5122 * Add experimental option FIPSSYMS to give all symbols in
5123 fipscanister.o and FIPS or fips prefix. This will avoid
5124 conflicts with future versions of OpenSSL. Add perl script
5125 util/fipsas.pl to preprocess assembly language source files
5126 and rename any affected symbols.
5130 * Add selftest checks and algorithm block of non-fips algorithms in
5131 FIPS mode. Remove DES2 from selftests.
5135 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5136 return internal method without any ENGINE dependencies. Add new
5137 tiny fips sign and verify functions.
5141 * New build option no-ec2m to disable characteristic 2 code.
5145 * New build option "fipscanisteronly". This only builds fipscanister.o
5146 and (currently) associated fips utilities. Uses the file Makefile.fips
5147 instead of Makefile.org as the prototype.
5151 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5152 Update fips_gcmtest to use IV generator.
5156 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5157 setting output buffer to NULL. The `*Final` function must be
5158 called although it will not retrieve any additional data. The tag
5159 can be set or retrieved with a ctrl. The IV length is by default 12
5160 bytes (96 bits) but can be set to an alternative value. If the IV
5161 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5166 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5167 underlying do_cipher function handles all cipher semantics itself
5168 including padding and finalisation. This is useful if (for example)
5169 an ENGINE cipher handles block padding itself. The behaviour of
5170 do_cipher is subtly changed if this flag is set: the return value
5171 is the number of characters written to the output buffer (zero is
5172 no longer an error code) or a negative error code. Also if the
5173 input buffer is NULL and length 0 finalisation should be performed.
5177 * If a candidate issuer certificate is already part of the constructed
5178 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5182 * Improve forward-security support: add functions
5184 void SSL_CTX_set_not_resumable_session_callback(
5185 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5186 void SSL_set_not_resumable_session_callback(
5187 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5189 for use by SSL/TLS servers; the callback function will be called whenever a
5190 new session is created, and gets to decide whether the session may be
5191 cached to make it resumable (return 0) or not (return 1). (As by the
5192 SSL/TLS protocol specifications, the session_id sent by the server will be
5193 empty to indicate that the session is not resumable; also, the server will
5194 not generate RFC 4507 (RFC 5077) session tickets.)
5196 A simple reasonable callback implementation is to return is_forward_secure.
5197 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5198 by the SSL/TLS server library, indicating whether it can provide forward
5201 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5203 * New -verify_name option in command line utilities to set verification
5208 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5209 Add CMAC pkey methods.
5213 * Experimental renegotiation in s_server -www mode. If the client
5214 browses /reneg connection is renegotiated. If /renegcert it is
5215 renegotiated requesting a certificate.
5219 * Add an "external" session cache for debugging purposes to s_server. This
5220 should help trace issues which normally are only apparent in deployed
5221 multi-process servers.
5225 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5226 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5227 BIO_set_cipher() and some obscure PEM functions were changed so they
5228 can now return an error. The RAND changes required a change to the
5229 RAND_METHOD structure.
5233 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5234 a gcc attribute to warn if the result of a function is ignored. This
5235 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5236 whose return value is often ignored.
5240 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5241 These allow SCTs (signed certificate timestamps) to be requested and
5242 validated when establishing a connection.
5244 *Rob Percival <robpercival@google.com>*
5249 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5251 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5252 used even when parsing explicit parameters, when loading a encoded key
5253 or calling `EC_GROUP_new_from_ecpkparameters()`/
5254 `EC_GROUP_new_from_ecparameters()`.
5255 This prevents bypass of security hardening and performance gains,
5256 especially for curves with specialized EC_METHODs.
5257 By default, if a key encoded with explicit parameters is loaded and later
5258 encoded, the output is still encoded with explicit parameters, even if
5259 internally a "named" EC_GROUP is used for computation.
5263 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5264 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5265 NULL. After this change, only the cofactor parameter can be NULL. It also
5266 does some minimal sanity checks on the passed order.
5271 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5272 An attack is simple, if the first CMS_recipientInfo is valid but the
5273 second CMS_recipientInfo is chosen ciphertext. If the second
5274 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5275 encryption key will be replaced by garbage, and the message cannot be
5276 decoded, but if the RSA decryption fails, the correct encryption key is
5277 used and the recipient will not notice the attack.
5278 As a work around for this potential attack the length of the decrypted
5279 key must be equal to the cipher default key length, in case the
5280 certificate is not given and all recipientInfo are tried out.
5281 The old behaviour can be re-enabled in the CMS code by setting the
5282 CMS_DEBUG_DECRYPT flag.
5287 * Document issue with installation paths in diverse Windows builds
5289 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5290 binaries and run-time config file.
5295 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5297 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5298 This changes the size when using the `genpkey` command when no size is given.
5299 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5300 generation commands to use 2048 bits by default.
5304 * Add FIPS support for Android Arm 64-bit
5306 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5307 Module in Version 2.0.10. For some reason, the corresponding target
5308 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5309 built with FIPS support on Android Arm 64-bit. This omission has been
5312 *Matthias St. Pierre*
5314 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5316 * 0-byte record padding oracle
5318 If an application encounters a fatal protocol error and then calls
5319 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5320 then OpenSSL can respond differently to the calling application if a 0 byte
5321 record is received with invalid padding compared to if a 0 byte record is
5322 received with an invalid MAC. If the application then behaves differently
5323 based on that in a way that is detectable to the remote peer, then this
5324 amounts to a padding oracle that could be used to decrypt data.
5326 In order for this to be exploitable "non-stitched" ciphersuites must be in
5327 use. Stitched ciphersuites are optimised implementations of certain
5328 commonly used ciphersuites. Also the application must call SSL_shutdown()
5329 twice even if a protocol error has occurred (applications should not do
5330 this but some do anyway).
5332 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5333 Aviram, with additional investigation by Steven Collison and Andrew
5334 Hourselt. It was reported to OpenSSL on 10th December 2018.
5339 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5343 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5345 * Microarchitecture timing vulnerability in ECC scalar multiplication
5347 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5348 shown to be vulnerable to a microarchitecture timing side channel attack.
5349 An attacker with sufficient access to mount local timing attacks during
5350 ECDSA signature generation could recover the private key.
5352 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5353 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5359 * Timing vulnerability in DSA signature generation
5361 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5362 timing side channel attack. An attacker could use variations in the signing
5363 algorithm to recover the private key.
5365 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5370 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5371 Module, accidentally introduced while backporting security fixes from the
5372 development branch and hindering the use of ECC in FIPS mode.
5376 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5378 * Client DoS due to large DH parameter
5380 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5381 malicious server can send a very large prime value to the client. This will
5382 cause the client to spend an unreasonably long period of time generating a
5383 key for this prime resulting in a hang until the client has finished. This
5384 could be exploited in a Denial Of Service attack.
5386 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5391 * Cache timing vulnerability in RSA Key Generation
5393 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5394 a cache timing side channel attack. An attacker with sufficient access to
5395 mount cache timing attacks during the RSA key generation process could
5396 recover the private key.
5398 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5399 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5404 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5405 parameter is no longer accepted, as it leads to a corrupt table. NULL
5406 pem_str is reserved for alias entries only.
5410 * Revert blinding in ECDSA sign and instead make problematic addition
5411 length-invariant. Switch even to fixed-length Montgomery multiplication.
5415 * Change generating and checking of primes so that the error rate of not
5416 being prime depends on the intended use based on the size of the input.
5417 For larger primes this will result in more rounds of Miller-Rabin.
5418 The maximal error rate for primes with more than 1080 bits is lowered
5421 *Kurt Roeckx, Annie Yousar*
5423 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5427 * Add blinding to ECDSA and DSA signatures to protect against side channel
5428 attacks discovered by Keegan Ryan (NCC Group).
5432 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5433 now allow empty (zero character) pass phrases.
5437 * Certificate time validation (X509_cmp_time) enforces stricter
5438 compliance with RFC 5280. Fractional seconds and timezone offsets
5439 are no longer allowed.
5443 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5445 * Constructed ASN.1 types with a recursive definition could exceed the stack
5447 Constructed ASN.1 types with a recursive definition (such as can be found
5448 in PKCS7) could eventually exceed the stack given malicious input with
5449 excessive recursion. This could result in a Denial Of Service attack. There
5450 are no such structures used within SSL/TLS that come from untrusted sources
5451 so this is considered safe.
5453 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5459 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5461 * Read/write after SSL object in error state
5463 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5464 mechanism. The intent was that if a fatal error occurred during a handshake
5465 then OpenSSL would move into the error state and would immediately fail if
5466 you attempted to continue the handshake. This works as designed for the
5467 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5468 SSL_connect()), however due to a bug it does not work correctly if
5469 SSL_read() or SSL_write() is called directly. In that scenario, if the
5470 handshake fails then a fatal error will be returned in the initial function
5471 call. If SSL_read()/SSL_write() is subsequently called by the application
5472 for the same SSL object then it will succeed and the data is passed without
5473 being decrypted/encrypted directly from the SSL/TLS record layer.
5475 In order to exploit this issue an application bug would have to be present
5476 that resulted in a call to SSL_read()/SSL_write() being issued after having
5477 already received a fatal error.
5479 This issue was reported to OpenSSL by David Benjamin (Google).
5484 * rsaz_1024_mul_avx2 overflow bug on x86_64
5486 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5487 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5488 Analysis suggests that attacks against RSA and DSA as a result of this
5489 defect would be very difficult to perform and are not believed likely.
5490 Attacks against DH1024 are considered just feasible, because most of the
5491 work necessary to deduce information about a private key may be performed
5492 offline. The amount of resources required for such an attack would be
5493 significant. However, for an attack on TLS to be meaningful, the server
5494 would have to share the DH1024 private key among multiple clients, which is
5495 no longer an option since CVE-2016-0701.
5497 This only affects processors that support the AVX2 but not ADX extensions
5498 like Intel Haswell (4th generation).
5500 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5501 was originally found via the OSS-Fuzz project.
5506 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5508 * bn_sqrx8x_internal carry bug on x86_64
5510 There is a carry propagating bug in the x86_64 Montgomery squaring
5511 procedure. No EC algorithms are affected. Analysis suggests that attacks
5512 against RSA and DSA as a result of this defect would be very difficult to
5513 perform and are not believed likely. Attacks against DH are considered just
5514 feasible (although very difficult) because most of the work necessary to
5515 deduce information about a private key may be performed offline. The amount
5516 of resources required for such an attack would be very significant and
5517 likely only accessible to a limited number of attackers. An attacker would
5518 additionally need online access to an unpatched system using the target
5519 private key in a scenario with persistent DH parameters and a private
5520 key that is shared between multiple clients.
5522 This only affects processors that support the BMI1, BMI2 and ADX extensions
5523 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5525 This issue was reported to OpenSSL by the OSS-Fuzz project.
5530 * Malformed X.509 IPAddressFamily could cause OOB read
5532 If an X.509 certificate has a malformed IPAddressFamily extension,
5533 OpenSSL could do a one-byte buffer overread. The most likely result
5534 would be an erroneous display of the certificate in text format.
5536 This issue was reported to OpenSSL by the OSS-Fuzz project.
5540 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5542 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5543 platform rather than 'mingw'.
5547 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5549 * Truncated packet could crash via OOB read
5551 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5552 cipher is being used, then a truncated packet can cause that host to
5553 perform an out-of-bounds read, usually resulting in a crash.
5555 This issue was reported to OpenSSL by Robert Święcki of Google.
5560 * BN_mod_exp may produce incorrect results on x86_64
5562 There is a carry propagating bug in the x86_64 Montgomery squaring
5563 procedure. No EC algorithms are affected. Analysis suggests that attacks
5564 against RSA and DSA as a result of this defect would be very difficult to
5565 perform and are not believed likely. Attacks against DH are considered just
5566 feasible (although very difficult) because most of the work necessary to
5567 deduce information about a private key may be performed offline. The amount
5568 of resources required for such an attack would be very significant and
5569 likely only accessible to a limited number of attackers. An attacker would
5570 additionally need online access to an unpatched system using the target
5571 private key in a scenario with persistent DH parameters and a private
5572 key that is shared between multiple clients. For example this can occur by
5573 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5574 similar to CVE-2015-3193 but must be treated as a separate problem.
5576 This issue was reported to OpenSSL by the OSS-Fuzz project.
5581 * Montgomery multiplication may produce incorrect results
5583 There is a carry propagating bug in the Broadwell-specific Montgomery
5584 multiplication procedure that handles input lengths divisible by, but
5585 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5586 and DH private keys are impossible. This is because the subroutine in
5587 question is not used in operations with the private key itself and an input
5588 of the attacker's direct choice. Otherwise the bug can manifest itself as
5589 transient authentication and key negotiation failures or reproducible
5590 erroneous outcome of public-key operations with specially crafted input.
5591 Among EC algorithms only Brainpool P-512 curves are affected and one
5592 presumably can attack ECDH key negotiation. Impact was not analyzed in
5593 detail, because pre-requisites for attack are considered unlikely. Namely
5594 multiple clients have to choose the curve in question and the server has to
5595 share the private key among them, neither of which is default behaviour.
5596 Even then only clients that chose the curve will be affected.
5598 This issue was publicly reported as transient failures and was not
5599 initially recognized as a security issue. Thanks to Richard Morgan for
5600 providing reproducible case.
5605 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5606 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5607 prevent issues where no progress is being made and the peer continually
5608 sends unrecognised record types, using up resources processing them.
5612 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5614 * Missing CRL sanity check
5616 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5617 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5618 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5620 This issue only affects the OpenSSL 1.0.2i
5625 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5627 * OCSP Status Request extension unbounded memory growth
5629 A malicious client can send an excessively large OCSP Status Request
5630 extension. If that client continually requests renegotiation, sending a
5631 large OCSP Status Request extension each time, then there will be unbounded
5632 memory growth on the server. This will eventually lead to a Denial Of
5633 Service attack through memory exhaustion. Servers with a default
5634 configuration are vulnerable even if they do not support OCSP. Builds using
5635 the "no-ocsp" build time option are not affected.
5637 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5642 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5645 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5651 * OOB write in MDC2_Update()
5653 An overflow can occur in MDC2_Update() either if called directly or
5654 through the EVP_DigestUpdate() function using MDC2. If an attacker
5655 is able to supply very large amounts of input data after a previous
5656 call to EVP_EncryptUpdate() with a partial block then a length check
5657 can overflow resulting in a heap corruption.
5659 The amount of data needed is comparable to SIZE_MAX which is impractical
5662 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5667 * Malformed SHA512 ticket DoS
5669 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5670 DoS attack where a malformed ticket will result in an OOB read which will
5673 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5674 a custom server callback and ticket lookup mechanism.
5676 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5681 * OOB write in BN_bn2dec()
5683 The function BN_bn2dec() does not check the return value of BN_div_word().
5684 This can cause an OOB write if an application uses this function with an
5685 overly large BIGNUM. This could be a problem if an overly large certificate
5686 or CRL is printed out from an untrusted source. TLS is not affected because
5687 record limits will reject an oversized certificate before it is parsed.
5689 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5694 * OOB read in TS_OBJ_print_bio()
5696 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5697 the total length the OID text representation would use and not the amount
5698 of data written. This will result in OOB reads when large OIDs are
5701 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5706 * Pointer arithmetic undefined behaviour
5708 Avoid some undefined pointer arithmetic
5710 A common idiom in the codebase is to check limits in the following manner:
5713 Where "p" points to some malloc'd data of SIZE bytes and
5716 "len" here could be from some externally supplied data (e.g. from a TLS
5719 The rules of C pointer arithmetic are such that "p + len" is only well
5720 defined where len <= SIZE. Therefore the above idiom is actually
5721 undefined behaviour.
5723 For example this could cause problems if some malloc implementation
5724 provides an address for "p" such that "p + len" actually overflows for
5725 values of len that are too big and therefore p + len < limit.
5727 This issue was reported to OpenSSL by Guido Vranken
5732 * Constant time flag not preserved in DSA signing
5734 Operations in the DSA signing algorithm should run in constant time in
5735 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5736 implementation means that a non-constant time codepath is followed for
5737 certain operations. This has been demonstrated through a cache-timing
5738 attack to be sufficient for an attacker to recover the private DSA key.
5740 This issue was reported by César Pereida (Aalto University), Billy Brumley
5741 (Tampere University of Technology), and Yuval Yarom (The University of
5742 Adelaide and NICTA).
5747 * DTLS buffered message DoS
5749 In a DTLS connection where handshake messages are delivered out-of-order
5750 those messages that OpenSSL is not yet ready to process will be buffered
5751 for later use. Under certain circumstances, a flaw in the logic means that
5752 those messages do not get removed from the buffer even though the handshake
5753 has been completed. An attacker could force up to approx. 15 messages to
5754 remain in the buffer when they are no longer required. These messages will
5755 be cleared when the DTLS connection is closed. The default maximum size for
5756 a message is 100k. Therefore, the attacker could force an additional 1500k
5757 to be consumed per connection. By opening many simultaneous connections an
5758 attacker could cause a DoS attack through memory exhaustion.
5760 This issue was reported to OpenSSL by Quan Luo.
5765 * DTLS replay protection DoS
5767 A flaw in the DTLS replay attack protection mechanism means that records
5768 that arrive for future epochs update the replay protection "window" before
5769 the MAC for the record has been validated. This could be exploited by an
5770 attacker by sending a record for the next epoch (which does not have to
5771 decrypt or have a valid MAC), with a very large sequence number. This means
5772 that all subsequent legitimate packets are dropped causing a denial of
5773 service for a specific DTLS connection.
5775 This issue was reported to OpenSSL by the OCAP audit team.
5780 * Certificate message OOB reads
5782 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5783 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5784 theoretical DoS risk but this has not been observed in practice on common
5787 The messages affected are client certificate, client certificate request
5788 and server certificate. As a result the attack can only be performed
5789 against a client or a server which enables client authentication.
5791 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5796 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5798 * Prevent padding oracle in AES-NI CBC MAC check
5800 A MITM attacker can use a padding oracle attack to decrypt traffic
5801 when the connection uses an AES CBC cipher and the server support
5804 This issue was introduced as part of the fix for Lucky 13 padding
5805 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5806 constant time by making sure that always the same bytes are read and
5807 compared against either the MAC or padding bytes. But it no longer
5808 checked that there was enough data to have both the MAC and padding
5811 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5815 * Fix EVP_EncodeUpdate overflow
5817 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5818 Base64 encoding of binary data. If an attacker is able to supply very large
5819 amounts of input data then a length check can overflow resulting in a heap
5822 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5823 the `PEM_write_bio*` family of functions. These are mainly used within the
5824 OpenSSL command line applications, so any application which processes data
5825 from an untrusted source and outputs it as a PEM file should be considered
5826 vulnerable to this issue. User applications that call these APIs directly
5827 with large amounts of untrusted data may also be vulnerable.
5829 This issue was reported by Guido Vranken.
5834 * Fix EVP_EncryptUpdate overflow
5836 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5837 is able to supply very large amounts of input data after a previous call to
5838 EVP_EncryptUpdate() with a partial block then a length check can overflow
5839 resulting in a heap corruption. Following an analysis of all OpenSSL
5840 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5841 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5842 the first called function after an EVP_EncryptInit(), and therefore that
5843 specific call must be safe. The second form is where the length passed to
5844 EVP_EncryptUpdate() can be seen from the code to be some small value and
5845 therefore there is no possibility of an overflow. Since all instances are
5846 one of these two forms, it is believed that there can be no overflows in
5847 internal code due to this problem. It should be noted that
5848 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5849 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5850 of these calls have also been analysed too and it is believed there are no
5851 instances in internal usage where an overflow could occur.
5853 This issue was reported by Guido Vranken.
5858 * Prevent ASN.1 BIO excessive memory allocation
5860 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5861 a short invalid encoding can cause allocation of large amounts of memory
5862 potentially consuming excessive resources or exhausting memory.
5864 Any application parsing untrusted data through d2i BIO functions is
5865 affected. The memory based functions such as d2i_X509() are *not* affected.
5866 Since the memory based functions are used by the TLS library, TLS
5867 applications are not affected.
5869 This issue was reported by Brian Carpenter.
5876 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5877 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5878 in arbitrary stack data being returned in the buffer.
5880 This issue was reported by Guido Vranken.
5885 * Modify behavior of ALPN to invoke callback after SNI/servername
5886 callback, such that updates to the SSL_CTX affect ALPN.
5890 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5895 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5896 methods are enabled and ssl2 is disabled the methods return NULL.
5900 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5902 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5903 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5904 provide any "EXPORT" or "LOW" strength ciphers.
5908 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5909 is by default disabled at build-time. Builds that are not configured with
5910 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5911 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5912 will need to explicitly call either of:
5914 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5916 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5918 as appropriate. Even if either of those is used, or the application
5919 explicitly uses the version-specific SSLv2_method() or its client and
5920 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5921 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5922 ciphers, and SSLv2 56-bit DES are no longer available.
5927 * Fix a double-free in DSA code
5929 A double free bug was discovered when OpenSSL parses malformed DSA private
5930 keys and could lead to a DoS attack or memory corruption for applications
5931 that receive DSA private keys from untrusted sources. This scenario is
5934 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5940 * Disable SRP fake user seed to address a server memory leak.
5942 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5944 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5945 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5946 was changed to ignore the "fake user" SRP seed, even if the seed
5949 Users should use SRP_VBASE_get1_by_user instead. Note that in
5950 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5951 also that even though configuring the SRP seed attempts to hide
5952 invalid usernames by continuing the handshake with fake
5953 credentials, this behaviour is not constant time and no strong
5954 guarantees are made that the handshake is indistinguishable from
5955 that of a valid user.
5960 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5962 In the BN_hex2bn function the number of hex digits is calculated using an
5963 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5964 large values of `i` this can result in `bn_expand` not allocating any
5965 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5966 field as NULL leading to a subsequent NULL ptr deref. For very large values
5967 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5968 In this case memory is allocated to the internal BIGNUM data field, but it
5969 is insufficiently sized leading to heap corruption. A similar issue exists
5970 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5971 is ever called by user applications with very large untrusted hex/dec data.
5972 This is anticipated to be a rare occurrence.
5974 All OpenSSL internal usage of these functions use data that is not expected
5975 to be untrusted, e.g. config file data or application command line
5976 arguments. If user developed applications generate config file data based
5977 on untrusted data then it is possible that this could also lead to security
5978 consequences. This is also anticipated to be rare.
5980 This issue was reported to OpenSSL by Guido Vranken.
5985 * Fix memory issues in `BIO_*printf` functions
5987 The internal `fmtstr` function used in processing a "%s" format string in
5988 the `BIO_*printf` functions could overflow while calculating the length of a
5989 string and cause an OOB read when printing very long strings.
5991 Additionally the internal `doapr_outch` function can attempt to write to an
5992 OOB memory location (at an offset from the NULL pointer) in the event of a
5993 memory allocation failure. In 1.0.2 and below this could be caused where
5994 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5995 could be in processing a very long "%s" format string. Memory leaks can
5998 The first issue may mask the second issue dependent on compiler behaviour.
5999 These problems could enable attacks where large amounts of untrusted data
6000 is passed to the `BIO_*printf` functions. If applications use these functions
6001 in this way then they could be vulnerable. OpenSSL itself uses these
6002 functions when printing out human-readable dumps of ASN.1 data. Therefore
6003 applications that print this data could be vulnerable if the data is from
6004 untrusted sources. OpenSSL command line applications could also be
6005 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6006 as command line arguments.
6008 Libssl is not considered directly vulnerable. Additionally certificates etc
6009 received via remote connections via libssl are also unlikely to be able to
6010 trigger these issues because of message size limits enforced within libssl.
6012 This issue was reported to OpenSSL Guido Vranken.
6017 * Side channel attack on modular exponentiation
6019 A side-channel attack was found which makes use of cache-bank conflicts on
6020 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6021 of RSA keys. The ability to exploit this issue is limited as it relies on
6022 an attacker who has control of code in a thread running on the same
6023 hyper-threaded core as the victim thread which is performing decryptions.
6025 This issue was reported to OpenSSL by Yuval Yarom, The University of
6026 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6027 Nadia Heninger, University of Pennsylvania with more information at
6028 <http://cachebleed.info>.
6033 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6034 if no keysize is specified with default_bits. This fixes an
6035 omission in an earlier change that changed all RSA/DSA key generation
6036 commands to use 2048 bits by default.
6040 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6042 * DH small subgroups
6044 Historically OpenSSL only ever generated DH parameters based on "safe"
6045 primes. More recently (in version 1.0.2) support was provided for
6046 generating X9.42 style parameter files such as those required for RFC 5114
6047 support. The primes used in such files may not be "safe". Where an
6048 application is using DH configured with parameters based on primes that are
6049 not "safe" then an attacker could use this fact to find a peer's private
6050 DH exponent. This attack requires that the attacker complete multiple
6051 handshakes in which the peer uses the same private DH exponent. For example
6052 this could be used to discover a TLS server's private DH exponent if it's
6053 reusing the private DH exponent or it's using a static DH ciphersuite.
6055 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6056 TLS. It is not on by default. If the option is not set then the server
6057 reuses the same private DH exponent for the life of the server process and
6058 would be vulnerable to this attack. It is believed that many popular
6059 applications do set this option and would therefore not be at risk.
6061 The fix for this issue adds an additional check where a "q" parameter is
6062 available (as is the case in X9.42 based parameters). This detects the
6063 only known attack, and is the only possible defense for static DH
6064 ciphersuites. This could have some performance impact.
6066 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6067 default and cannot be disabled. This could have some performance impact.
6069 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6074 * SSLv2 doesn't block disabled ciphers
6076 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6077 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6078 been disabled, provided that the SSLv2 protocol was not also disabled via
6081 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6082 and Sebastian Schinzel.
6087 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6089 * BN_mod_exp may produce incorrect results on x86_64
6091 There is a carry propagating bug in the x86_64 Montgomery squaring
6092 procedure. No EC algorithms are affected. Analysis suggests that attacks
6093 against RSA and DSA as a result of this defect would be very difficult to
6094 perform and are not believed likely. Attacks against DH are considered just
6095 feasible (although very difficult) because most of the work necessary to
6096 deduce information about a private key may be performed offline. The amount
6097 of resources required for such an attack would be very significant and
6098 likely only accessible to a limited number of attackers. An attacker would
6099 additionally need online access to an unpatched system using the target
6100 private key in a scenario with persistent DH parameters and a private
6101 key that is shared between multiple clients. For example this can occur by
6102 default in OpenSSL DHE based SSL/TLS ciphersuites.
6104 This issue was reported to OpenSSL by Hanno Böck.
6109 * Certificate verify crash with missing PSS parameter
6111 The signature verification routines will crash with a NULL pointer
6112 dereference if presented with an ASN.1 signature using the RSA PSS
6113 algorithm and absent mask generation function parameter. Since these
6114 routines are used to verify certificate signature algorithms this can be
6115 used to crash any certificate verification operation and exploited in a
6116 DoS attack. Any application which performs certificate verification is
6117 vulnerable including OpenSSL clients and servers which enable client
6120 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6125 * X509_ATTRIBUTE memory leak
6127 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6128 memory. This structure is used by the PKCS#7 and CMS routines so any
6129 application which reads PKCS#7 or CMS data from untrusted sources is
6130 affected. SSL/TLS is not affected.
6132 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6138 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6139 This changes the decoding behaviour for some invalid messages,
6140 though the change is mostly in the more lenient direction, and
6141 legacy behaviour is preserved as much as possible.
6145 * In DSA_generate_parameters_ex, if the provided seed is too short,
6148 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6150 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6152 * Alternate chains certificate forgery
6154 During certificate verification, OpenSSL will attempt to find an
6155 alternative certificate chain if the first attempt to build such a chain
6156 fails. An error in the implementation of this logic can mean that an
6157 attacker could cause certain checks on untrusted certificates to be
6158 bypassed, such as the CA flag, enabling them to use a valid leaf
6159 certificate to act as a CA and "issue" an invalid certificate.
6161 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6166 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6168 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6169 incompatibility in the handling of HMAC. The previous ABI has now been
6174 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6176 * Malformed ECParameters causes infinite loop
6178 When processing an ECParameters structure OpenSSL enters an infinite loop
6179 if the curve specified is over a specially malformed binary polynomial
6182 This can be used to perform denial of service against any
6183 system which processes public keys, certificate requests or
6184 certificates. This includes TLS clients and TLS servers with
6185 client authentication enabled.
6187 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6192 * Exploitable out-of-bounds read in X509_cmp_time
6194 X509_cmp_time does not properly check the length of the ASN1_TIME
6195 string and can read a few bytes out of bounds. In addition,
6196 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6199 An attacker can use this to craft malformed certificates and CRLs of
6200 various sizes and potentially cause a segmentation fault, resulting in
6201 a DoS on applications that verify certificates or CRLs. TLS clients
6202 that verify CRLs are affected. TLS clients and servers with client
6203 authentication enabled may be affected if they use custom verification
6206 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6207 independently by Hanno Böck.
6212 * PKCS7 crash with missing EnvelopedContent
6214 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6215 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6216 with missing content and trigger a NULL pointer dereference on parsing.
6218 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6219 structures from untrusted sources are affected. OpenSSL clients and
6220 servers are not affected.
6222 This issue was reported to OpenSSL by Michal Zalewski (Google).
6227 * CMS verify infinite loop with unknown hash function
6229 When verifying a signedData message the CMS code can enter an infinite loop
6230 if presented with an unknown hash function OID. This can be used to perform
6231 denial of service against any system which verifies signedData messages using
6233 This issue was reported to OpenSSL by Johannes Bauer.
6238 * Race condition handling NewSessionTicket
6240 If a NewSessionTicket is received by a multi-threaded client when attempting to
6241 reuse a previous ticket then a race condition can occur potentially leading to
6242 a double free of the ticket data.
6247 * Only support 256-bit or stronger elliptic curves with the
6248 'ecdh_auto' setting (server) or by default (client). Of supported
6249 curves, prefer P-256 (both).
6253 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6255 * ClientHello sigalgs DoS fix
6257 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6258 invalid signature algorithms extension a NULL pointer dereference will
6259 occur. This can be exploited in a DoS attack against the server.
6261 This issue was was reported to OpenSSL by David Ramos of Stanford
6265 *Stephen Henson and Matt Caswell*
6267 * Multiblock corrupted pointer fix
6269 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6270 feature only applies on 64 bit x86 architecture platforms that support AES
6271 NI instructions. A defect in the implementation of "multiblock" can cause
6272 OpenSSL's internal write buffer to become incorrectly set to NULL when
6273 using non-blocking IO. Typically, when the user application is using a
6274 socket BIO for writing, this will only result in a failed connection.
6275 However if some other BIO is used then it is likely that a segmentation
6276 fault will be triggered, thus enabling a potential DoS attack.
6278 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6283 * Segmentation fault in DTLSv1_listen fix
6285 The DTLSv1_listen function is intended to be stateless and processes the
6286 initial ClientHello from many peers. It is common for user code to loop
6287 over the call to DTLSv1_listen until a valid ClientHello is received with
6288 an associated cookie. A defect in the implementation of DTLSv1_listen means
6289 that state is preserved in the SSL object from one invocation to the next
6290 that can lead to a segmentation fault. Errors processing the initial
6291 ClientHello can trigger this scenario. An example of such an error could be
6292 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6295 This issue was reported to OpenSSL by Per Allansson.
6300 * Segmentation fault in ASN1_TYPE_cmp fix
6302 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6303 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6304 certificate signature algorithm consistency this can be used to crash any
6305 certificate verification operation and exploited in a DoS attack. Any
6306 application which performs certificate verification is vulnerable including
6307 OpenSSL clients and servers which enable client authentication.
6312 * Segmentation fault for invalid PSS parameters fix
6314 The signature verification routines will crash with a NULL pointer
6315 dereference if presented with an ASN.1 signature using the RSA PSS
6316 algorithm and invalid parameters. Since these routines are used to verify
6317 certificate signature algorithms this can be used to crash any
6318 certificate verification operation and exploited in a DoS attack. Any
6319 application which performs certificate verification is vulnerable including
6320 OpenSSL clients and servers which enable client authentication.
6322 This issue was was reported to OpenSSL by Brian Carpenter.
6327 * ASN.1 structure reuse memory corruption fix
6329 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6330 memory corruption via an invalid write. Such reuse is and has been
6331 strongly discouraged and is believed to be rare.
6333 Applications that parse structures containing CHOICE or ANY DEFINED BY
6334 components may be affected. Certificate parsing (d2i_X509 and related
6335 functions) are however not affected. OpenSSL clients and servers are
6341 * PKCS7 NULL pointer dereferences fix
6343 The PKCS#7 parsing code does not handle missing outer ContentInfo
6344 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6345 missing content and trigger a NULL pointer dereference on parsing.
6347 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6348 otherwise parse PKCS#7 structures from untrusted sources are
6349 affected. OpenSSL clients and servers are not affected.
6351 This issue was reported to OpenSSL by Michal Zalewski (Google).
6356 * DoS via reachable assert in SSLv2 servers fix
6358 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6359 servers that both support SSLv2 and enable export cipher suites by sending
6360 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6362 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6363 (OpenSSL development team).
6368 * Empty CKE with client auth and DHE fix
6370 If client auth is used then a server can seg fault in the event of a DHE
6371 ciphersuite being selected and a zero length ClientKeyExchange message
6372 being sent by the client. This could be exploited in a DoS attack.
6377 * Handshake with unseeded PRNG fix
6379 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6380 with an unseeded PRNG. The conditions are:
6381 - The client is on a platform where the PRNG has not been seeded
6382 automatically, and the user has not seeded manually
6383 - A protocol specific client method version has been used (i.e. not
6384 SSL_client_methodv23)
6385 - A ciphersuite is used that does not require additional random data from
6386 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6388 If the handshake succeeds then the client random that has been used will
6389 have been generated from a PRNG with insufficient entropy and therefore the
6390 output may be predictable.
6392 For example using the following command with an unseeded openssl will
6393 succeed on an unpatched platform:
6395 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6400 * Use After Free following d2i_ECPrivatekey error fix
6402 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6403 could cause a use after free condition. This, in turn, could cause a double
6404 free in several private key parsing functions (such as d2i_PrivateKey
6405 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6406 for applications that receive EC private keys from untrusted
6407 sources. This scenario is considered rare.
6409 This issue was discovered by the BoringSSL project and fixed in their
6415 * X509_to_X509_REQ NULL pointer deref fix
6417 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6418 the certificate key is invalid. This function is rarely used in practice.
6420 This issue was discovered by Brian Carpenter.
6425 * Removed the export ciphers from the DEFAULT ciphers
6429 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6431 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6432 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6433 So far those who have to target multiple platforms would compromise
6434 and argue that binary targeting say ARMv5 would still execute on
6435 ARMv8. "Universal" build resolves this compromise by providing
6436 near-optimal performance even on newer platforms.
6440 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6441 (other platforms pending).
6443 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6445 * Add support for the SignedCertificateTimestampList certificate and
6446 OCSP response extensions from RFC6962.
6450 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6451 for corner cases. (Certain input points at infinity could lead to
6452 bogus results, with non-infinity inputs mapped to infinity too.)
6456 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6457 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6458 common cases are optimized and there still is room for further
6459 improvements. Vector Permutation AES for Altivec is also added.
6463 * Add support for little-endian ppc64 Linux target.
6465 *Marcelo Cerri (IBM)*
6467 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6468 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6469 are optimized and there still is room for further improvements.
6470 Both 32- and 64-bit modes are supported.
6472 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6474 * Improved ARMv7 NEON support.
6478 * Support for SPARC Architecture 2011 crypto extensions, first
6479 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6480 SHA256/512, MD5, GHASH and modular exponentiation.
6482 *Andy Polyakov, David Miller*
6484 * Accelerated modular exponentiation for Intel processors, a.k.a.
6487 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6489 * Support for new and upcoming Intel processors, including AVX2,
6490 BMI and SHA ISA extensions. This includes additional "stitched"
6491 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6494 This work was sponsored by Intel Corp.
6498 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6499 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6500 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6504 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6505 this fixes a limitation in previous versions of OpenSSL.
6509 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6510 MGF1 digest and OAEP label.
6514 * Add EVP support for key wrapping algorithms, to avoid problems with
6515 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6516 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6517 algorithms and include tests cases.
6521 * Add functions to allocate and set the fields of an ECDSA_METHOD
6524 *Douglas E. Engert, Steve Henson*
6526 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6527 difference in days and seconds between two tm or ASN1_TIME structures.
6531 * Add -rev test option to s_server to just reverse order of characters
6532 received by client and send back to server. Also prints an abbreviated
6533 summary of the connection parameters.
6537 * New option -brief for s_client and s_server to print out a brief summary
6538 of connection parameters.
6542 * Add callbacks for arbitrary TLS extensions.
6544 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6546 * New option -crl_download in several openssl utilities to download CRLs
6547 from CRLDP extension in certificates.
6551 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6555 * New function X509_CRL_diff to generate a delta CRL from the difference
6556 of two full CRLs. Add support to "crl" utility.
6560 * New functions to set lookup_crls function and to retrieve
6561 X509_STORE from X509_STORE_CTX.
6565 * Print out deprecated issuer and subject unique ID fields in
6570 * Extend OCSP I/O functions so they can be used for simple general purpose
6571 HTTP as well as OCSP. New wrapper function which can be used to download
6572 CRLs using the OCSP API.
6576 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6580 * `SSL_CONF*` functions. These provide a common framework for application
6581 configuration using configuration files or command lines.
6585 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6586 message callback and prints the results. Needs compile time option
6587 "enable-ssl-trace". New options to s_client and s_server to enable
6592 * New ctrl and macro to retrieve supported points extensions.
6593 Print out extension in s_server and s_client.
6597 * New functions to retrieve certificate signature and signature
6602 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6607 * New Suite B modes for TLS code. These use and enforce the requirements
6608 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6609 only use Suite B curves. The Suite B modes can be set by using the
6610 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6614 * New chain verification flags for Suite B levels of security. Check
6615 algorithms are acceptable when flags are set in X509_verify_cert.
6619 * Make tls1_check_chain return a set of flags indicating checks passed
6620 by a certificate chain. Add additional tests to handle client
6621 certificates: checks for matching certificate type and issuer name
6626 * If an attempt is made to use a signature algorithm not in the peer
6627 preference list abort the handshake. If client has no suitable
6628 signature algorithms in response to a certificate request do not
6629 use the certificate.
6633 * If server EC tmp key is not in client preference list abort handshake.
6637 * Add support for certificate stores in CERT structure. This makes it
6638 possible to have different stores per SSL structure or one store in
6639 the parent SSL_CTX. Include distinct stores for certificate chain
6640 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6641 to build and store a certificate chain in CERT structure: returning
6642 an error if the chain cannot be built: this will allow applications
6643 to test if a chain is correctly configured.
6645 Note: if the CERT based stores are not set then the parent SSL_CTX
6646 store is used to retain compatibility with existing behaviour.
6650 * New function ssl_set_client_disabled to set a ciphersuite disabled
6651 mask based on the current session, check mask when sending client
6652 hello and checking the requested ciphersuite.
6656 * New ctrls to retrieve and set certificate types in a certificate
6657 request message. Print out received values in s_client. If certificate
6658 types is not set with custom values set sensible values based on
6659 supported signature algorithms.
6663 * Support for distinct client and server supported signature algorithms.
6667 * Add certificate callback. If set this is called whenever a certificate
6668 is required by client or server. An application can decide which
6669 certificate chain to present based on arbitrary criteria: for example
6670 supported signature algorithms. Add very simple example to s_server.
6671 This fixes many of the problems and restrictions of the existing client
6672 certificate callback: for example you can now clear an existing
6673 certificate and specify the whole chain.
6677 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6678 the certificate can be used for (if anything). Set valid_flags field
6679 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6680 to have similar checks in it.
6682 Add new "cert_flags" field to CERT structure and include a "strict mode".
6683 This enforces some TLS certificate requirements (such as only permitting
6684 certificate signature algorithms contained in the supported algorithms
6685 extension) which some implementations ignore: this option should be used
6686 with caution as it could cause interoperability issues.
6690 * Update and tidy signature algorithm extension processing. Work out
6691 shared signature algorithms based on preferences and peer algorithms
6692 and print them out in s_client and s_server. Abort handshake if no
6693 shared signature algorithms.
6697 * Add new functions to allow customised supported signature algorithms
6698 for SSL and SSL_CTX structures. Add options to s_client and s_server
6703 * New function SSL_certs_clear() to delete all references to certificates
6704 from an SSL structure. Before this once a certificate had been added
6705 it couldn't be removed.
6709 * Integrate hostname, email address and IP address checking with certificate
6710 verification. New verify options supporting checking in openssl utility.
6714 * Fixes and wildcard matching support to hostname and email checking
6715 functions. Add manual page.
6717 *Florian Weimer (Red Hat Product Security Team)*
6719 * New functions to check a hostname email or IP address against a
6720 certificate. Add options x509 utility to print results of checks against
6725 * Fix OCSP checking.
6727 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6729 * Initial experimental support for explicitly trusted non-root CAs.
6730 OpenSSL still tries to build a complete chain to a root but if an
6731 intermediate CA has a trust setting included that is used. The first
6732 setting is used: whether to trust (e.g., -addtrust option to the x509
6737 * Add -trusted_first option which attempts to find certificates in the
6738 trusted store even if an untrusted chain is also supplied.
6742 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6743 platform support for Linux and Android.
6747 * Support for linux-x32, ILP32 environment in x86_64 framework.
6751 * Experimental multi-implementation support for FIPS capable OpenSSL.
6752 When in FIPS mode the approved implementations are used as normal,
6753 when not in FIPS mode the internal unapproved versions are used instead.
6754 This means that the FIPS capable OpenSSL isn't forced to use the
6755 (often lower performance) FIPS implementations outside FIPS mode.
6759 * Transparently support X9.42 DH parameters when calling
6760 PEM_read_bio_DHparameters. This means existing applications can handle
6761 the new parameter format automatically.
6765 * Initial experimental support for X9.42 DH parameter format: mainly
6766 to support use of 'q' parameter for RFC5114 parameters.
6770 * Add DH parameters from RFC5114 including test data to dhtest.
6774 * Support for automatic EC temporary key parameter selection. If enabled
6775 the most preferred EC parameters are automatically used instead of
6776 hardcoded fixed parameters. Now a server just has to call:
6777 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6778 support ECDH and use the most appropriate parameters.
6782 * Enhance and tidy EC curve and point format TLS extension code. Use
6783 static structures instead of allocation if default values are used.
6784 New ctrls to set curves we wish to support and to retrieve shared curves.
6785 Print out shared curves in s_server. New options to s_server and s_client
6786 to set list of supported curves.
6790 * New ctrls to retrieve supported signature algorithms and
6791 supported curve values as an array of NIDs. Extend openssl utility
6792 to print out received values.
6796 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6797 between NIDs and the more common NIST names such as "P-256". Enhance
6798 ecparam utility and ECC method to recognise the NIST names for curves.
6802 * Enhance SSL/TLS certificate chain handling to support different
6803 chains for each certificate instead of one chain in the parent SSL_CTX.
6807 * Support for fixed DH ciphersuite client authentication: where both
6808 server and client use DH certificates with common parameters.
6812 * Support for fixed DH ciphersuites: those requiring DH server
6817 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6819 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6820 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6821 X509_CINF_get_signature were reverted post internal team review.
6826 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6828 * OCSP Status Request extension unbounded memory growth
6830 A malicious client can send an excessively large OCSP Status Request
6831 extension. If that client continually requests renegotiation, sending a
6832 large OCSP Status Request extension each time, then there will be unbounded
6833 memory growth on the server. This will eventually lead to a Denial Of
6834 Service attack through memory exhaustion. Servers with a default
6835 configuration are vulnerable even if they do not support OCSP. Builds using
6836 the "no-ocsp" build time option are not affected.
6838 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6843 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6846 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6852 * OOB write in MDC2_Update()
6854 An overflow can occur in MDC2_Update() either if called directly or
6855 through the EVP_DigestUpdate() function using MDC2. If an attacker
6856 is able to supply very large amounts of input data after a previous
6857 call to EVP_EncryptUpdate() with a partial block then a length check
6858 can overflow resulting in a heap corruption.
6860 The amount of data needed is comparable to SIZE_MAX which is impractical
6863 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6868 * Malformed SHA512 ticket DoS
6870 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6871 DoS attack where a malformed ticket will result in an OOB read which will
6874 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6875 a custom server callback and ticket lookup mechanism.
6877 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6882 * OOB write in BN_bn2dec()
6884 The function BN_bn2dec() does not check the return value of BN_div_word().
6885 This can cause an OOB write if an application uses this function with an
6886 overly large BIGNUM. This could be a problem if an overly large certificate
6887 or CRL is printed out from an untrusted source. TLS is not affected because
6888 record limits will reject an oversized certificate before it is parsed.
6890 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6895 * OOB read in TS_OBJ_print_bio()
6897 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6898 the total length the OID text representation would use and not the amount
6899 of data written. This will result in OOB reads when large OIDs are
6902 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6907 * Pointer arithmetic undefined behaviour
6909 Avoid some undefined pointer arithmetic
6911 A common idiom in the codebase is to check limits in the following manner:
6914 Where "p" points to some malloc'd data of SIZE bytes and
6917 "len" here could be from some externally supplied data (e.g. from a TLS
6920 The rules of C pointer arithmetic are such that "p + len" is only well
6921 defined where len <= SIZE. Therefore, the above idiom is actually
6922 undefined behaviour.
6924 For example this could cause problems if some malloc implementation
6925 provides an address for "p" such that "p + len" actually overflows for
6926 values of len that are too big and therefore p + len < limit.
6928 This issue was reported to OpenSSL by Guido Vranken
6933 * Constant time flag not preserved in DSA signing
6935 Operations in the DSA signing algorithm should run in constant time in
6936 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6937 implementation means that a non-constant time codepath is followed for
6938 certain operations. This has been demonstrated through a cache-timing
6939 attack to be sufficient for an attacker to recover the private DSA key.
6941 This issue was reported by César Pereida (Aalto University), Billy Brumley
6942 (Tampere University of Technology), and Yuval Yarom (The University of
6943 Adelaide and NICTA).
6948 * DTLS buffered message DoS
6950 In a DTLS connection where handshake messages are delivered out-of-order
6951 those messages that OpenSSL is not yet ready to process will be buffered
6952 for later use. Under certain circumstances, a flaw in the logic means that
6953 those messages do not get removed from the buffer even though the handshake
6954 has been completed. An attacker could force up to approx. 15 messages to
6955 remain in the buffer when they are no longer required. These messages will
6956 be cleared when the DTLS connection is closed. The default maximum size for
6957 a message is 100k. Therefore, the attacker could force an additional 1500k
6958 to be consumed per connection. By opening many simultaneous connections an
6959 attacker could cause a DoS attack through memory exhaustion.
6961 This issue was reported to OpenSSL by Quan Luo.
6966 * DTLS replay protection DoS
6968 A flaw in the DTLS replay attack protection mechanism means that records
6969 that arrive for future epochs update the replay protection "window" before
6970 the MAC for the record has been validated. This could be exploited by an
6971 attacker by sending a record for the next epoch (which does not have to
6972 decrypt or have a valid MAC), with a very large sequence number. This means
6973 that all subsequent legitimate packets are dropped causing a denial of
6974 service for a specific DTLS connection.
6976 This issue was reported to OpenSSL by the OCAP audit team.
6981 * Certificate message OOB reads
6983 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6984 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6985 theoretical DoS risk but this has not been observed in practice on common
6988 The messages affected are client certificate, client certificate request
6989 and server certificate. As a result the attack can only be performed
6990 against a client or a server which enables client authentication.
6992 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6997 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6999 * Prevent padding oracle in AES-NI CBC MAC check
7001 A MITM attacker can use a padding oracle attack to decrypt traffic
7002 when the connection uses an AES CBC cipher and the server support
7005 This issue was introduced as part of the fix for Lucky 13 padding
7006 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7007 constant time by making sure that always the same bytes are read and
7008 compared against either the MAC or padding bytes. But it no longer
7009 checked that there was enough data to have both the MAC and padding
7012 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7017 * Fix EVP_EncodeUpdate overflow
7019 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7020 Base64 encoding of binary data. If an attacker is able to supply very large
7021 amounts of input data then a length check can overflow resulting in a heap
7024 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7025 the `PEM_write_bio*` family of functions. These are mainly used within the
7026 OpenSSL command line applications, so any application which processes data
7027 from an untrusted source and outputs it as a PEM file should be considered
7028 vulnerable to this issue. User applications that call these APIs directly
7029 with large amounts of untrusted data may also be vulnerable.
7031 This issue was reported by Guido Vranken.
7036 * Fix EVP_EncryptUpdate overflow
7038 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7039 is able to supply very large amounts of input data after a previous call to
7040 EVP_EncryptUpdate() with a partial block then a length check can overflow
7041 resulting in a heap corruption. Following an analysis of all OpenSSL
7042 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7043 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7044 the first called function after an EVP_EncryptInit(), and therefore that
7045 specific call must be safe. The second form is where the length passed to
7046 EVP_EncryptUpdate() can be seen from the code to be some small value and
7047 therefore there is no possibility of an overflow. Since all instances are
7048 one of these two forms, it is believed that there can be no overflows in
7049 internal code due to this problem. It should be noted that
7050 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7051 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7052 of these calls have also been analysed too and it is believed there are no
7053 instances in internal usage where an overflow could occur.
7055 This issue was reported by Guido Vranken.
7060 * Prevent ASN.1 BIO excessive memory allocation
7062 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7063 a short invalid encoding can casuse allocation of large amounts of memory
7064 potentially consuming excessive resources or exhausting memory.
7066 Any application parsing untrusted data through d2i BIO functions is
7067 affected. The memory based functions such as d2i_X509() are *not* affected.
7068 Since the memory based functions are used by the TLS library, TLS
7069 applications are not affected.
7071 This issue was reported by Brian Carpenter.
7078 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7079 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7080 in arbitrary stack data being returned in the buffer.
7082 This issue was reported by Guido Vranken.
7087 * Modify behavior of ALPN to invoke callback after SNI/servername
7088 callback, such that updates to the SSL_CTX affect ALPN.
7092 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7097 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7098 methods are enabled and ssl2 is disabled the methods return NULL.
7102 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7104 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7105 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7106 provide any "EXPORT" or "LOW" strength ciphers.
7110 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7111 is by default disabled at build-time. Builds that are not configured with
7112 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7113 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7114 will need to explicitly call either of:
7116 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7118 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7120 as appropriate. Even if either of those is used, or the application
7121 explicitly uses the version-specific SSLv2_method() or its client and
7122 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7123 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7124 ciphers, and SSLv2 56-bit DES are no longer available.
7129 * Fix a double-free in DSA code
7131 A double free bug was discovered when OpenSSL parses malformed DSA private
7132 keys and could lead to a DoS attack or memory corruption for applications
7133 that receive DSA private keys from untrusted sources. This scenario is
7136 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7142 * Disable SRP fake user seed to address a server memory leak.
7144 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7146 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7147 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7148 was changed to ignore the "fake user" SRP seed, even if the seed
7151 Users should use SRP_VBASE_get1_by_user instead. Note that in
7152 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7153 also that even though configuring the SRP seed attempts to hide
7154 invalid usernames by continuing the handshake with fake
7155 credentials, this behaviour is not constant time and no strong
7156 guarantees are made that the handshake is indistinguishable from
7157 that of a valid user.
7162 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7164 In the BN_hex2bn function the number of hex digits is calculated using an
7165 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7166 large values of `i` this can result in `bn_expand` not allocating any
7167 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7168 field as NULL leading to a subsequent NULL ptr deref. For very large values
7169 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7170 In this case memory is allocated to the internal BIGNUM data field, but it
7171 is insufficiently sized leading to heap corruption. A similar issue exists
7172 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7173 is ever called by user applications with very large untrusted hex/dec data.
7174 This is anticipated to be a rare occurrence.
7176 All OpenSSL internal usage of these functions use data that is not expected
7177 to be untrusted, e.g. config file data or application command line
7178 arguments. If user developed applications generate config file data based
7179 on untrusted data then it is possible that this could also lead to security
7180 consequences. This is also anticipated to be rare.
7182 This issue was reported to OpenSSL by Guido Vranken.
7187 * Fix memory issues in `BIO_*printf` functions
7189 The internal `fmtstr` function used in processing a "%s" format string in
7190 the `BIO_*printf` functions could overflow while calculating the length of a
7191 string and cause an OOB read when printing very long strings.
7193 Additionally the internal `doapr_outch` function can attempt to write to an
7194 OOB memory location (at an offset from the NULL pointer) in the event of a
7195 memory allocation failure. In 1.0.2 and below this could be caused where
7196 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7197 could be in processing a very long "%s" format string. Memory leaks can
7200 The first issue may mask the second issue dependent on compiler behaviour.
7201 These problems could enable attacks where large amounts of untrusted data
7202 is passed to the `BIO_*printf` functions. If applications use these functions
7203 in this way then they could be vulnerable. OpenSSL itself uses these
7204 functions when printing out human-readable dumps of ASN.1 data. Therefore
7205 applications that print this data could be vulnerable if the data is from
7206 untrusted sources. OpenSSL command line applications could also be
7207 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7208 as command line arguments.
7210 Libssl is not considered directly vulnerable. Additionally certificates etc
7211 received via remote connections via libssl are also unlikely to be able to
7212 trigger these issues because of message size limits enforced within libssl.
7214 This issue was reported to OpenSSL Guido Vranken.
7219 * Side channel attack on modular exponentiation
7221 A side-channel attack was found which makes use of cache-bank conflicts on
7222 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7223 of RSA keys. The ability to exploit this issue is limited as it relies on
7224 an attacker who has control of code in a thread running on the same
7225 hyper-threaded core as the victim thread which is performing decryptions.
7227 This issue was reported to OpenSSL by Yuval Yarom, The University of
7228 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7229 Nadia Heninger, University of Pennsylvania with more information at
7230 <http://cachebleed.info>.
7235 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7236 if no keysize is specified with default_bits. This fixes an
7237 omission in an earlier change that changed all RSA/DSA key generation
7238 commands to use 2048 bits by default.
7242 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7244 * Protection for DH small subgroup attacks
7246 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7247 switched on by default and cannot be disabled. This could have some
7252 * SSLv2 doesn't block disabled ciphers
7254 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7255 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7256 been disabled, provided that the SSLv2 protocol was not also disabled via
7259 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7260 and Sebastian Schinzel.
7265 * Reject DH handshakes with parameters shorter than 1024 bits.
7269 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7271 * Certificate verify crash with missing PSS parameter
7273 The signature verification routines will crash with a NULL pointer
7274 dereference if presented with an ASN.1 signature using the RSA PSS
7275 algorithm and absent mask generation function parameter. Since these
7276 routines are used to verify certificate signature algorithms this can be
7277 used to crash any certificate verification operation and exploited in a
7278 DoS attack. Any application which performs certificate verification is
7279 vulnerable including OpenSSL clients and servers which enable client
7282 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7287 * X509_ATTRIBUTE memory leak
7289 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7290 memory. This structure is used by the PKCS#7 and CMS routines so any
7291 application which reads PKCS#7 or CMS data from untrusted sources is
7292 affected. SSL/TLS is not affected.
7294 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7300 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7301 This changes the decoding behaviour for some invalid messages,
7302 though the change is mostly in the more lenient direction, and
7303 legacy behaviour is preserved as much as possible.
7307 * In DSA_generate_parameters_ex, if the provided seed is too short,
7308 use a random seed, as already documented.
7310 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7312 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7314 * Alternate chains certificate forgery
7316 During certificate verfification, OpenSSL will attempt to find an
7317 alternative certificate chain if the first attempt to build such a chain
7318 fails. An error in the implementation of this logic can mean that an
7319 attacker could cause certain checks on untrusted certificates to be
7320 bypassed, such as the CA flag, enabling them to use a valid leaf
7321 certificate to act as a CA and "issue" an invalid certificate.
7323 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7329 * Race condition handling PSK identify hint
7331 If PSK identity hints are received by a multi-threaded client then
7332 the values are wrongly updated in the parent SSL_CTX structure. This can
7333 result in a race condition potentially leading to a double free of the
7339 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7341 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7342 incompatibility in the handling of HMAC. The previous ABI has now been
7345 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7347 * Malformed ECParameters causes infinite loop
7349 When processing an ECParameters structure OpenSSL enters an infinite loop
7350 if the curve specified is over a specially malformed binary polynomial
7353 This can be used to perform denial of service against any
7354 system which processes public keys, certificate requests or
7355 certificates. This includes TLS clients and TLS servers with
7356 client authentication enabled.
7358 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7363 * Exploitable out-of-bounds read in X509_cmp_time
7365 X509_cmp_time does not properly check the length of the ASN1_TIME
7366 string and can read a few bytes out of bounds. In addition,
7367 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7370 An attacker can use this to craft malformed certificates and CRLs of
7371 various sizes and potentially cause a segmentation fault, resulting in
7372 a DoS on applications that verify certificates or CRLs. TLS clients
7373 that verify CRLs are affected. TLS clients and servers with client
7374 authentication enabled may be affected if they use custom verification
7377 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7378 independently by Hanno Böck.
7383 * PKCS7 crash with missing EnvelopedContent
7385 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7386 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7387 with missing content and trigger a NULL pointer dereference on parsing.
7389 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7390 structures from untrusted sources are affected. OpenSSL clients and
7391 servers are not affected.
7393 This issue was reported to OpenSSL by Michal Zalewski (Google).
7398 * CMS verify infinite loop with unknown hash function
7400 When verifying a signedData message the CMS code can enter an infinite loop
7401 if presented with an unknown hash function OID. This can be used to perform
7402 denial of service against any system which verifies signedData messages using
7404 This issue was reported to OpenSSL by Johannes Bauer.
7409 * Race condition handling NewSessionTicket
7411 If a NewSessionTicket is received by a multi-threaded client when attempting to
7412 reuse a previous ticket then a race condition can occur potentially leading to
7413 a double free of the ticket data.
7418 * Reject DH handshakes with parameters shorter than 768 bits.
7420 *Kurt Roeckx and Emilia Kasper*
7422 * dhparam: generate 2048-bit parameters by default.
7424 *Kurt Roeckx and Emilia Kasper*
7426 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7428 * Segmentation fault in ASN1_TYPE_cmp fix
7430 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7431 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7432 certificate signature algorithm consistency this can be used to crash any
7433 certificate verification operation and exploited in a DoS attack. Any
7434 application which performs certificate verification is vulnerable including
7435 OpenSSL clients and servers which enable client authentication.
7440 * ASN.1 structure reuse memory corruption fix
7442 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7443 memory corruption via an invalid write. Such reuse is and has been
7444 strongly discouraged and is believed to be rare.
7446 Applications that parse structures containing CHOICE or ANY DEFINED BY
7447 components may be affected. Certificate parsing (d2i_X509 and related
7448 functions) are however not affected. OpenSSL clients and servers are
7454 * PKCS7 NULL pointer dereferences fix
7456 The PKCS#7 parsing code does not handle missing outer ContentInfo
7457 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7458 missing content and trigger a NULL pointer dereference on parsing.
7460 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7461 otherwise parse PKCS#7 structures from untrusted sources are
7462 affected. OpenSSL clients and servers are not affected.
7464 This issue was reported to OpenSSL by Michal Zalewski (Google).
7469 * DoS via reachable assert in SSLv2 servers fix
7471 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7472 servers that both support SSLv2 and enable export cipher suites by sending
7473 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7475 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7476 (OpenSSL development team).
7481 * Use After Free following d2i_ECPrivatekey error fix
7483 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7484 could cause a use after free condition. This, in turn, could cause a double
7485 free in several private key parsing functions (such as d2i_PrivateKey
7486 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7487 for applications that receive EC private keys from untrusted
7488 sources. This scenario is considered rare.
7490 This issue was discovered by the BoringSSL project and fixed in their
7496 * X509_to_X509_REQ NULL pointer deref fix
7498 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7499 the certificate key is invalid. This function is rarely used in practice.
7501 This issue was discovered by Brian Carpenter.
7506 * Removed the export ciphers from the DEFAULT ciphers
7510 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7512 * Build fixes for the Windows and OpenVMS platforms
7514 *Matt Caswell and Richard Levitte*
7516 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7518 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7519 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7520 dereference. This could lead to a Denial Of Service attack. Thanks to
7521 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7526 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7527 dtls1_buffer_record function under certain conditions. In particular this
7528 could occur if an attacker sent repeated DTLS records with the same
7529 sequence number but for the next epoch. The memory leak could be exploited
7530 by an attacker in a Denial of Service attack through memory exhaustion.
7531 Thanks to Chris Mueller for reporting this issue.
7536 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7537 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7538 method would be set to NULL which could later result in a NULL pointer
7539 dereference. Thanks to Frank Schmirler for reporting this issue.
7544 * Abort handshake if server key exchange message is omitted for ephemeral
7547 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7548 reporting this issue.
7553 * Remove non-export ephemeral RSA code on client and server. This code
7554 violated the TLS standard by allowing the use of temporary RSA keys in
7555 non-export ciphersuites and could be used by a server to effectively
7556 downgrade the RSA key length used to a value smaller than the server
7557 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7558 INRIA or reporting this issue.
7563 * Fixed issue where DH client certificates are accepted without verification.
7564 An OpenSSL server will accept a DH certificate for client authentication
7565 without the certificate verify message. This effectively allows a client to
7566 authenticate without the use of a private key. This only affects servers
7567 which trust a client certificate authority which issues certificates
7568 containing DH keys: these are extremely rare and hardly ever encountered.
7569 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7575 * Ensure that the session ID context of an SSL is updated when its
7576 SSL_CTX is updated via SSL_set_SSL_CTX.
7578 The session ID context is typically set from the parent SSL_CTX,
7579 and can vary with the CTX.
7583 * Fix various certificate fingerprint issues.
7585 By using non-DER or invalid encodings outside the signed portion of a
7586 certificate the fingerprint can be changed without breaking the signature.
7587 Although no details of the signed portion of the certificate can be changed
7588 this can cause problems with some applications: e.g. those using the
7589 certificate fingerprint for blacklists.
7591 1. Reject signatures with non zero unused bits.
7593 If the BIT STRING containing the signature has non zero unused bits reject
7594 the signature. All current signature algorithms require zero unused bits.
7596 2. Check certificate algorithm consistency.
7598 Check the AlgorithmIdentifier inside TBS matches the one in the
7599 certificate signature. NB: this will result in signature failure
7600 errors for some broken certificates.
7602 Thanks to Konrad Kraszewski from Google for reporting this issue.
7604 3. Check DSA/ECDSA signatures use DER.
7606 Re-encode DSA/ECDSA signatures and compare with the original received
7607 signature. Return an error if there is a mismatch.
7609 This will reject various cases including garbage after signature
7610 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7611 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7612 (negative or with leading zeroes).
7614 Further analysis was conducted and fixes were developed by Stephen Henson
7615 of the OpenSSL core team.
7621 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7622 results on some platforms, including x86_64. This bug occurs at random
7623 with a very low probability, and is not known to be exploitable in any
7624 way, though its exact impact is difficult to determine. Thanks to Pieter
7625 Wuille (Blockstream) who reported this issue and also suggested an initial
7626 fix. Further analysis was conducted by the OpenSSL development team and
7627 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7628 the OpenSSL core team.
7633 * Do not resume sessions on the server if the negotiated protocol
7634 version does not match the session's version. Resuming with a different
7635 version, while not strictly forbidden by the RFC, is of questionable
7636 sanity and breaks all known clients.
7638 *David Benjamin, Emilia Käsper*
7640 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7641 early CCS messages during renegotiation. (Note that because
7642 renegotiation is encrypted, this early CCS was not exploitable.)
7646 * Tighten client-side session ticket handling during renegotiation:
7647 ensure that the client only accepts a session ticket if the server sends
7648 the extension anew in the ServerHello. Previously, a TLS client would
7649 reuse the old extension state and thus accept a session ticket if one was
7650 announced in the initial ServerHello.
7652 Similarly, ensure that the client requires a session ticket if one
7653 was advertised in the ServerHello. Previously, a TLS client would
7654 ignore a missing NewSessionTicket message.
7658 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7662 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7663 sends a carefully crafted handshake message, to cause OpenSSL to fail
7664 to free up to 64k of memory causing a memory leak. This could be
7665 exploited in a Denial Of Service attack. This issue affects OpenSSL
7666 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7667 whether SRTP is used or configured. Implementations of OpenSSL that
7668 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7670 The fix was developed by the OpenSSL team.
7675 * Session Ticket Memory Leak.
7677 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7678 integrity of that ticket is first verified. In the event of a session
7679 ticket integrity check failing, OpenSSL will fail to free memory
7680 causing a memory leak. By sending a large number of invalid session
7681 tickets an attacker could exploit this issue in a Denial Of Service
7687 * Build option no-ssl3 is incomplete.
7689 When OpenSSL is configured with "no-ssl3" as a build option, servers
7690 could accept and complete an SSL 3.0 handshake, and clients could be
7691 configured to send them.
7694 *Akamai and the OpenSSL team*
7696 * Add support for TLS_FALLBACK_SCSV.
7697 Client applications doing fallback retries should call
7698 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7701 *Adam Langley, Bodo Moeller*
7703 * Add additional DigestInfo checks.
7705 Re-encode DigestInto in DER and check against the original when
7706 verifying RSA signature: this will reject any improperly encoded
7707 DigestInfo structures.
7709 Note: this is a precautionary measure and no attacks are currently known.
7713 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7715 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7716 SRP code can be overrun an internal buffer. Add sanity check that
7717 g, A, B < N to SRP code.
7719 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7720 Group for discovering this issue.
7725 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7726 TLS 1.0 instead of higher protocol versions when the ClientHello message
7727 is badly fragmented. This allows a man-in-the-middle attacker to force a
7728 downgrade to TLS 1.0 even if both the server and the client support a
7729 higher protocol version, by modifying the client's TLS records.
7731 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7732 researching this issue.
7737 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7738 to a denial of service attack. A malicious server can crash the client
7739 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7740 ciphersuite and sending carefully crafted handshake messages.
7742 Thanks to Felix Gröbert (Google) for discovering and researching this
7748 * By sending carefully crafted DTLS packets an attacker could cause openssl
7749 to leak memory. This can be exploited through a Denial of Service attack.
7750 Thanks to Adam Langley for discovering and researching this issue.
7755 * An attacker can force openssl to consume large amounts of memory whilst
7756 processing DTLS handshake messages. This can be exploited through a
7757 Denial of Service attack.
7758 Thanks to Adam Langley for discovering and researching this issue.
7763 * An attacker can force an error condition which causes openssl to crash
7764 whilst processing DTLS packets due to memory being freed twice. This
7765 can be exploited through a Denial of Service attack.
7766 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7772 * If a multithreaded client connects to a malicious server using a resumed
7773 session and the server sends an ec point format extension it could write
7774 up to 255 bytes to freed memory.
7776 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7782 * A malicious server can crash an OpenSSL client with a null pointer
7783 dereference (read) by specifying an SRP ciphersuite even though it was not
7784 properly negotiated with the client. This can be exploited through a
7785 Denial of Service attack.
7787 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7788 discovering and researching this issue.
7793 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7794 X509_name_oneline, X509_name_print_ex et al. to leak some information
7795 from the stack. Applications may be affected if they echo pretty printing
7796 output to the attacker.
7798 Thanks to Ivan Fratric (Google) for discovering this issue.
7801 *Emilia Käsper, and Steve Henson*
7803 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7804 for corner cases. (Certain input points at infinity could lead to
7805 bogus results, with non-infinity inputs mapped to infinity too.)
7809 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7811 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7812 handshake can force the use of weak keying material in OpenSSL
7813 SSL/TLS clients and servers.
7815 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7816 researching this issue. ([CVE-2014-0224])
7818 *KIKUCHI Masashi, Steve Henson*
7820 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7821 OpenSSL DTLS client the code can be made to recurse eventually crashing
7824 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7827 *Imre Rad, Steve Henson*
7829 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7830 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7831 client or server. This is potentially exploitable to run arbitrary
7832 code on a vulnerable client or server.
7834 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7836 *Jüri Aedla, Steve Henson*
7838 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7839 are subject to a denial of service attack.
7841 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7842 this issue. ([CVE-2014-3470])
7844 *Felix Gröbert, Ivan Fratric, Steve Henson*
7846 * Harmonize version and its documentation. -f flag is used to display
7849 *mancha <mancha1@zoho.com>*
7851 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7852 in i2d_ECPrivateKey.
7854 *mancha <mancha1@zoho.com>*
7856 * Fix some double frees. These are not thought to be exploitable.
7858 *mancha <mancha1@zoho.com>*
7860 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7862 * A missing bounds check in the handling of the TLS heartbeat extension
7863 can be used to reveal up to 64k of memory to a connected client or
7866 Thanks for Neel Mehta of Google Security for discovering this bug and to
7867 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7868 preparing the fix ([CVE-2014-0160])
7870 *Adam Langley, Bodo Moeller*
7872 * Fix for the attack described in the paper "Recovering OpenSSL
7873 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7874 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7875 <http://eprint.iacr.org/2014/140>
7877 Thanks to Yuval Yarom and Naomi Benger for discovering this
7878 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7880 *Yuval Yarom and Naomi Benger*
7882 * TLS pad extension: draft-agl-tls-padding-03
7884 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7885 TLS client Hello record length value would otherwise be > 255 and
7886 less that 512 pad with a dummy extension containing zeroes so it
7887 is at least 512 bytes long.
7889 *Adam Langley, Steve Henson*
7891 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7893 * Fix for TLS record tampering bug. A carefully crafted invalid
7894 handshake could crash OpenSSL with a NULL pointer exception.
7895 Thanks to Anton Johansson for reporting this issues.
7898 * Keep original DTLS digest and encryption contexts in retransmission
7899 structures so we can use the previous session parameters if they need
7900 to be resent. ([CVE-2013-6450])
7904 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7905 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7906 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7907 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7908 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7909 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7911 *Rob Stradling, Adam Langley*
7913 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7915 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7916 supporting platforms or when small records were transferred.
7918 *Andy Polyakov, Steve Henson*
7920 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7922 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7924 This addresses the flaw in CBC record processing discovered by
7925 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7926 at: <http://www.isg.rhul.ac.uk/tls/>
7928 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7929 Security Group at Royal Holloway, University of London
7930 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7931 Emilia Käsper for the initial patch.
7934 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7936 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7937 ciphersuites which can be exploited in a denial of service attack.
7938 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7939 and detecting this bug and to Wolfgang Ettlinger
7940 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7945 * Return an error when checking OCSP signatures when key is NULL.
7946 This fixes a DoS attack. ([CVE-2013-0166])
7950 * Make openssl verify return errors.
7952 *Chris Palmer <palmer@google.com> and Ben Laurie*
7954 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7955 the right response is stapled. Also change SSL_get_certificate()
7956 so it returns the certificate actually sent.
7957 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7959 *Rob Stradling <rob.stradling@comodo.com>*
7961 * Fix possible deadlock when decoding public keys.
7965 * Don't use TLS 1.0 record version number in initial client hello
7970 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7972 * Sanity check record length before skipping explicit IV in TLS
7973 1.2, 1.1 and DTLS to fix DoS attack.
7975 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7976 fuzzing as a service testing platform.
7981 * Initialise tkeylen properly when encrypting CMS messages.
7982 Thanks to Solar Designer of Openwall for reporting this issue.
7986 * In FIPS mode don't try to use composite ciphers as they are not
7991 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7993 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7994 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7995 mean any application compiled against OpenSSL 1.0.0 headers setting
7996 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7997 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7998 0x10000000L Any application which was previously compiled against
7999 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8000 will need to be recompiled as a result. Letting be results in
8001 inability to disable specifically TLS 1.1 and in client context,
8002 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8006 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8007 disable just protocol X, but all protocols above X *if* there are
8008 protocols *below* X still enabled. In more practical terms it means
8009 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8010 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8011 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8016 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8018 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8019 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8020 in CRYPTO_realloc_clean.
8022 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8023 issue and to Adam Langley <agl@chromium.org> for fixing it.
8026 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8028 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8032 * Workarounds for some broken servers that "hang" if a client hello
8033 record length exceeds 255 bytes.
8035 1. Do not use record version number > TLS 1.0 in initial client
8036 hello: some (but not all) hanging servers will now work.
8037 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8038 the number of ciphers sent in the client hello. This should be
8039 set to an even number, such as 50, for example by passing:
8040 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8041 Most broken servers should now work.
8042 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8043 TLS 1.2 client support entirely.
8047 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8051 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8053 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8054 STRING form instead of a DigestInfo.
8058 * The format used for MDC2 RSA signatures is inconsistent between EVP
8059 and the RSA_sign/RSA_verify functions. This was made more apparent when
8060 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8061 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8062 the correct format in RSA_verify so both forms transparently work.
8066 * Some servers which support TLS 1.0 can choke if we initially indicate
8067 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8068 encrypted premaster secret. As a workaround use the maximum permitted
8069 client version in client hello, this should keep such servers happy
8070 and still work with previous versions of OpenSSL.
8074 * Add support for TLS/DTLS heartbeats.
8076 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8078 * Add support for SCTP.
8080 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8082 * Improved PRNG seeding for VOS.
8084 *Paul Green <Paul.Green@stratus.com>*
8086 * Extensive assembler packs updates, most notably:
8088 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8089 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8090 - x86_64: bit-sliced AES implementation;
8091 - ARM: NEON support, contemporary platforms optimizations;
8092 - s390x: z196 support;
8093 - `*`: GHASH and GF(2^m) multiplication implementations;
8097 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8098 (removal of unnecessary code)
8100 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8102 * Add TLS key material exporter from RFC 5705.
8106 * Add DTLS-SRTP negotiation from RFC 5764.
8110 * Add Next Protocol Negotiation,
8111 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8112 disabled with a no-npn flag to config or Configure. Code donated
8115 *Adam Langley <agl@google.com> and Ben Laurie*
8117 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8118 NIST-P256, NIST-P521, with constant-time single point multiplication on
8119 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8120 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8121 Code made available under Apache License version 2.0.
8123 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8124 line to include this in your build of OpenSSL, and run "make depend" (or
8125 "make update"). This enables the following EC_METHODs:
8127 EC_GFp_nistp224_method()
8128 EC_GFp_nistp256_method()
8129 EC_GFp_nistp521_method()
8131 EC_GROUP_new_by_curve_name() will automatically use these (while
8132 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8135 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8137 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8138 all platforms. Move ssize_t definition from e_os.h to the public
8139 header file e_os2.h as it now appears in public header file cms.h
8143 * New -sigopt option to the ca, req and x509 utilities. Additional
8144 signature parameters can be passed using this option and in
8149 * Add RSA PSS signing function. This will generate and set the
8150 appropriate AlgorithmIdentifiers for PSS based on those in the
8151 corresponding EVP_MD_CTX structure. No application support yet.
8155 * Support for companion algorithm specific ASN1 signing routines.
8156 New function ASN1_item_sign_ctx() signs a pre-initialised
8157 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8158 the appropriate parameters.
8162 * Add new algorithm specific ASN1 verification initialisation function
8163 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8164 handling will be the same no matter what EVP_PKEY_METHOD is used.
8165 Add a PSS handler to support verification of PSS signatures: checked
8166 against a number of sample certificates.
8170 * Add signature printing for PSS. Add PSS OIDs.
8172 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8174 * Add algorithm specific signature printing. An individual ASN1 method
8175 can now print out signatures instead of the standard hex dump.
8177 More complex signatures (e.g. PSS) can print out more meaningful
8178 information. Include DSA version that prints out the signature
8183 * Password based recipient info support for CMS library: implementing
8188 * Split password based encryption into PBES2 and PBKDF2 functions. This
8189 neatly separates the code into cipher and PBE sections and is required
8190 for some algorithms that split PBES2 into separate pieces (such as
8191 password based CMS).
8195 * Session-handling fixes:
8196 - Fix handling of connections that are resuming with a session ID,
8197 but also support Session Tickets.
8198 - Fix a bug that suppressed issuing of a new ticket if the client
8199 presented a ticket with an expired session.
8200 - Try to set the ticket lifetime hint to something reasonable.
8201 - Make tickets shorter by excluding irrelevant information.
8202 - On the client side, don't ignore renewed tickets.
8204 *Adam Langley, Bodo Moeller (Google)*
8206 * Fix PSK session representation.
8210 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8212 This work was sponsored by Intel.
8216 * Add GCM support to TLS library. Some custom code is needed to split
8217 the IV between the fixed (from PRF) and explicit (from TLS record)
8218 portions. This adds all GCM ciphersuites supported by RFC5288 and
8219 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8220 add a special AESGCM string for GCM only.
8224 * Expand range of ctrls for AES GCM. Permit setting invocation
8225 field on decrypt and retrieval of invocation field only on encrypt.
8229 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8230 As required by RFC5289 these ciphersuites cannot be used if for
8231 versions of TLS earlier than 1.2.
8235 * For FIPS capable OpenSSL interpret a NULL default public key method
8236 as unset and return the appropriate default but do *not* set the default.
8237 This means we can return the appropriate method in applications that
8238 switch between FIPS and non-FIPS modes.
8242 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8243 ENGINE is used then we cannot handle that in the FIPS module so we
8244 keep original code iff non-FIPS operations are allowed.
8248 * Add -attime option to openssl utilities.
8250 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8252 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8256 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8257 FIPS EC methods unconditionally for now.
8261 * New build option no-ec2m to disable characteristic 2 code.
8265 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8266 all cases can be covered as some introduce binary incompatibilities.
8270 * Redirect RSA operations to FIPS module including keygen,
8271 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8275 * Add similar low-level API blocking to ciphers.
8279 * low-level digest APIs are not approved in FIPS mode: any attempt
8280 to use these will cause a fatal error. Applications that *really* want
8281 to use them can use the `private_*` version instead.
8285 * Redirect cipher operations to FIPS module for FIPS builds.
8289 * Redirect digest operations to FIPS module for FIPS builds.
8293 * Update build system to add "fips" flag which will link in fipscanister.o
8294 for static and shared library builds embedding a signature if needed.
8298 * Output TLS supported curves in preference order instead of numerical
8299 order. This is currently hardcoded for the highest order curves first.
8300 This should be configurable so applications can judge speed vs strength.
8304 * Add TLS v1.2 server support for client authentication.
8308 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8313 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8314 FIPS modules versions.
8318 * Add TLS v1.2 client side support for client authentication. Keep cache
8319 of handshake records longer as we don't know the hash algorithm to use
8320 until after the certificate request message is received.
8324 * Initial TLS v1.2 client support. Add a default signature algorithms
8325 extension including all the algorithms we support. Parse new signature
8326 format in client key exchange. Relax some ECC signing restrictions for
8327 TLS v1.2 as indicated in RFC5246.
8331 * Add server support for TLS v1.2 signature algorithms extension. Switch
8332 to new signature format when needed using client digest preference.
8333 All server ciphersuites should now work correctly in TLS v1.2. No client
8334 support yet and no support for client certificates.
8338 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8339 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8340 ciphersuites. At present only RSA key exchange ciphersuites work with
8341 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8342 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8343 and version checking.
8347 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8348 with this defined it will not be affected by any changes to ssl internal
8349 structures. Add several utility functions to allow openssl application
8350 to work with OPENSSL_NO_SSL_INTERN defined.
8354 * A long standing patch to add support for SRP from EdelWeb (Peter
8355 Sylvester and Christophe Renou) was integrated.
8356 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8357 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8360 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8364 * Permit abbreviated handshakes when renegotiating using the function
8365 SSL_renegotiate_abbreviated().
8367 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8369 * Add call to ENGINE_register_all_complete() to
8370 ENGINE_load_builtin_engines(), so some implementations get used
8371 automatically instead of needing explicit application support.
8375 * Add support for TLS key exporter as described in RFC5705.
8377 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8379 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8380 a few changes are required:
8382 Add SSL_OP_NO_TLSv1_1 flag.
8383 Add TLSv1_1 methods.
8384 Update version checking logic to handle version 1.1.
8385 Add explicit IV handling (ported from DTLS code).
8386 Add command line options to s_client/s_server.
8393 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8395 * X509_ATTRIBUTE memory leak
8397 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8398 memory. This structure is used by the PKCS#7 and CMS routines so any
8399 application which reads PKCS#7 or CMS data from untrusted sources is
8400 affected. SSL/TLS is not affected.
8402 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8408 * Race condition handling PSK identify hint
8410 If PSK identity hints are received by a multi-threaded client then
8411 the values are wrongly updated in the parent SSL_CTX structure. This can
8412 result in a race condition potentially leading to a double free of the
8418 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8420 * Malformed ECParameters causes infinite loop
8422 When processing an ECParameters structure OpenSSL enters an infinite loop
8423 if the curve specified is over a specially malformed binary polynomial
8426 This can be used to perform denial of service against any
8427 system which processes public keys, certificate requests or
8428 certificates. This includes TLS clients and TLS servers with
8429 client authentication enabled.
8431 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8436 * Exploitable out-of-bounds read in X509_cmp_time
8438 X509_cmp_time does not properly check the length of the ASN1_TIME
8439 string and can read a few bytes out of bounds. In addition,
8440 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8443 An attacker can use this to craft malformed certificates and CRLs of
8444 various sizes and potentially cause a segmentation fault, resulting in
8445 a DoS on applications that verify certificates or CRLs. TLS clients
8446 that verify CRLs are affected. TLS clients and servers with client
8447 authentication enabled may be affected if they use custom verification
8450 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8451 independently by Hanno Böck.
8456 * PKCS7 crash with missing EnvelopedContent
8458 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8459 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8460 with missing content and trigger a NULL pointer dereference on parsing.
8462 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8463 structures from untrusted sources are affected. OpenSSL clients and
8464 servers are not affected.
8466 This issue was reported to OpenSSL by Michal Zalewski (Google).
8471 * CMS verify infinite loop with unknown hash function
8473 When verifying a signedData message the CMS code can enter an infinite loop
8474 if presented with an unknown hash function OID. This can be used to perform
8475 denial of service against any system which verifies signedData messages using
8477 This issue was reported to OpenSSL by Johannes Bauer.
8482 * Race condition handling NewSessionTicket
8484 If a NewSessionTicket is received by a multi-threaded client when attempting to
8485 reuse a previous ticket then a race condition can occur potentially leading to
8486 a double free of the ticket data.
8491 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8493 * Segmentation fault in ASN1_TYPE_cmp fix
8495 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8496 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8497 certificate signature algorithm consistency this can be used to crash any
8498 certificate verification operation and exploited in a DoS attack. Any
8499 application which performs certificate verification is vulnerable including
8500 OpenSSL clients and servers which enable client authentication.
8505 * ASN.1 structure reuse memory corruption fix
8507 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8508 memory corruption via an invalid write. Such reuse is and has been
8509 strongly discouraged and is believed to be rare.
8511 Applications that parse structures containing CHOICE or ANY DEFINED BY
8512 components may be affected. Certificate parsing (d2i_X509 and related
8513 functions) are however not affected. OpenSSL clients and servers are
8519 * PKCS7 NULL pointer dereferences fix
8521 The PKCS#7 parsing code does not handle missing outer ContentInfo
8522 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8523 missing content and trigger a NULL pointer dereference on parsing.
8525 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8526 otherwise parse PKCS#7 structures from untrusted sources are
8527 affected. OpenSSL clients and servers are not affected.
8529 This issue was reported to OpenSSL by Michal Zalewski (Google).
8534 * DoS via reachable assert in SSLv2 servers fix
8536 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8537 servers that both support SSLv2 and enable export cipher suites by sending
8538 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8540 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8541 (OpenSSL development team).
8546 * Use After Free following d2i_ECPrivatekey error fix
8548 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8549 could cause a use after free condition. This, in turn, could cause a double
8550 free in several private key parsing functions (such as d2i_PrivateKey
8551 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8552 for applications that receive EC private keys from untrusted
8553 sources. This scenario is considered rare.
8555 This issue was discovered by the BoringSSL project and fixed in their
8561 * X509_to_X509_REQ NULL pointer deref fix
8563 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8564 the certificate key is invalid. This function is rarely used in practice.
8566 This issue was discovered by Brian Carpenter.
8571 * Removed the export ciphers from the DEFAULT ciphers
8575 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8577 * Build fixes for the Windows and OpenVMS platforms
8579 *Matt Caswell and Richard Levitte*
8581 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8583 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8584 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8585 dereference. This could lead to a Denial Of Service attack. Thanks to
8586 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8591 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8592 dtls1_buffer_record function under certain conditions. In particular this
8593 could occur if an attacker sent repeated DTLS records with the same
8594 sequence number but for the next epoch. The memory leak could be exploited
8595 by an attacker in a Denial of Service attack through memory exhaustion.
8596 Thanks to Chris Mueller for reporting this issue.
8601 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8602 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8603 method would be set to NULL which could later result in a NULL pointer
8604 dereference. Thanks to Frank Schmirler for reporting this issue.
8609 * Abort handshake if server key exchange message is omitted for ephemeral
8612 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8613 reporting this issue.
8618 * Remove non-export ephemeral RSA code on client and server. This code
8619 violated the TLS standard by allowing the use of temporary RSA keys in
8620 non-export ciphersuites and could be used by a server to effectively
8621 downgrade the RSA key length used to a value smaller than the server
8622 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8623 INRIA or reporting this issue.
8628 * Fixed issue where DH client certificates are accepted without verification.
8629 An OpenSSL server will accept a DH certificate for client authentication
8630 without the certificate verify message. This effectively allows a client to
8631 authenticate without the use of a private key. This only affects servers
8632 which trust a client certificate authority which issues certificates
8633 containing DH keys: these are extremely rare and hardly ever encountered.
8634 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8640 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8641 results on some platforms, including x86_64. This bug occurs at random
8642 with a very low probability, and is not known to be exploitable in any
8643 way, though its exact impact is difficult to determine. Thanks to Pieter
8644 Wuille (Blockstream) who reported this issue and also suggested an initial
8645 fix. Further analysis was conducted by the OpenSSL development team and
8646 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8647 the OpenSSL core team.
8652 * Fix various certificate fingerprint issues.
8654 By using non-DER or invalid encodings outside the signed portion of a
8655 certificate the fingerprint can be changed without breaking the signature.
8656 Although no details of the signed portion of the certificate can be changed
8657 this can cause problems with some applications: e.g. those using the
8658 certificate fingerprint for blacklists.
8660 1. Reject signatures with non zero unused bits.
8662 If the BIT STRING containing the signature has non zero unused bits reject
8663 the signature. All current signature algorithms require zero unused bits.
8665 2. Check certificate algorithm consistency.
8667 Check the AlgorithmIdentifier inside TBS matches the one in the
8668 certificate signature. NB: this will result in signature failure
8669 errors for some broken certificates.
8671 Thanks to Konrad Kraszewski from Google for reporting this issue.
8673 3. Check DSA/ECDSA signatures use DER.
8675 Reencode DSA/ECDSA signatures and compare with the original received
8676 signature. Return an error if there is a mismatch.
8678 This will reject various cases including garbage after signature
8679 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8680 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8681 (negative or with leading zeroes).
8683 Further analysis was conducted and fixes were developed by Stephen Henson
8684 of the OpenSSL core team.
8690 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8692 * Session Ticket Memory Leak.
8694 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8695 integrity of that ticket is first verified. In the event of a session
8696 ticket integrity check failing, OpenSSL will fail to free memory
8697 causing a memory leak. By sending a large number of invalid session
8698 tickets an attacker could exploit this issue in a Denial Of Service
8704 * Build option no-ssl3 is incomplete.
8706 When OpenSSL is configured with "no-ssl3" as a build option, servers
8707 could accept and complete an SSL 3.0 handshake, and clients could be
8708 configured to send them.
8711 *Akamai and the OpenSSL team*
8713 * Add support for TLS_FALLBACK_SCSV.
8714 Client applications doing fallback retries should call
8715 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8718 *Adam Langley, Bodo Moeller*
8720 * Add additional DigestInfo checks.
8722 Reencode DigestInto in DER and check against the original when
8723 verifying RSA signature: this will reject any improperly encoded
8724 DigestInfo structures.
8726 Note: this is a precautionary measure and no attacks are currently known.
8730 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8732 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8733 to a denial of service attack. A malicious server can crash the client
8734 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8735 ciphersuite and sending carefully crafted handshake messages.
8737 Thanks to Felix Gröbert (Google) for discovering and researching this
8743 * By sending carefully crafted DTLS packets an attacker could cause openssl
8744 to leak memory. This can be exploited through a Denial of Service attack.
8745 Thanks to Adam Langley for discovering and researching this issue.
8750 * An attacker can force openssl to consume large amounts of memory whilst
8751 processing DTLS handshake messages. This can be exploited through a
8752 Denial of Service attack.
8753 Thanks to Adam Langley for discovering and researching this issue.
8758 * An attacker can force an error condition which causes openssl to crash
8759 whilst processing DTLS packets due to memory being freed twice. This
8760 can be exploited through a Denial of Service attack.
8761 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8767 * If a multithreaded client connects to a malicious server using a resumed
8768 session and the server sends an ec point format extension it could write
8769 up to 255 bytes to freed memory.
8771 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8777 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8778 X509_name_oneline, X509_name_print_ex et al. to leak some information
8779 from the stack. Applications may be affected if they echo pretty printing
8780 output to the attacker.
8782 Thanks to Ivan Fratric (Google) for discovering this issue.
8785 *Emilia Käsper, and Steve Henson*
8787 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8788 for corner cases. (Certain input points at infinity could lead to
8789 bogus results, with non-infinity inputs mapped to infinity too.)
8793 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8795 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8796 handshake can force the use of weak keying material in OpenSSL
8797 SSL/TLS clients and servers.
8799 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8800 researching this issue. ([CVE-2014-0224])
8802 *KIKUCHI Masashi, Steve Henson*
8804 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8805 OpenSSL DTLS client the code can be made to recurse eventually crashing
8808 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8811 *Imre Rad, Steve Henson*
8813 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8814 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8815 client or server. This is potentially exploitable to run arbitrary
8816 code on a vulnerable client or server.
8818 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8820 *Jüri Aedla, Steve Henson*
8822 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8823 are subject to a denial of service attack.
8825 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8826 this issue. ([CVE-2014-3470])
8828 *Felix Gröbert, Ivan Fratric, Steve Henson*
8830 * Harmonize version and its documentation. -f flag is used to display
8833 *mancha <mancha1@zoho.com>*
8835 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8836 in i2d_ECPrivateKey.
8838 *mancha <mancha1@zoho.com>*
8840 * Fix some double frees. These are not thought to be exploitable.
8842 *mancha <mancha1@zoho.com>*
8844 * Fix for the attack described in the paper "Recovering OpenSSL
8845 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8846 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8847 <http://eprint.iacr.org/2014/140>
8849 Thanks to Yuval Yarom and Naomi Benger for discovering this
8850 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8852 *Yuval Yarom and Naomi Benger*
8854 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8856 * Keep original DTLS digest and encryption contexts in retransmission
8857 structures so we can use the previous session parameters if they need
8858 to be resent. ([CVE-2013-6450])
8862 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8863 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8864 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8865 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8866 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8867 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8869 *Rob Stradling, Adam Langley*
8871 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8873 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8875 This addresses the flaw in CBC record processing discovered by
8876 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8877 at: <http://www.isg.rhul.ac.uk/tls/>
8879 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8880 Security Group at Royal Holloway, University of London
8881 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8882 Emilia Käsper for the initial patch.
8885 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8887 * Return an error when checking OCSP signatures when key is NULL.
8888 This fixes a DoS attack. ([CVE-2013-0166])
8892 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8893 the right response is stapled. Also change SSL_get_certificate()
8894 so it returns the certificate actually sent.
8895 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8896 (This is a backport)
8898 *Rob Stradling <rob.stradling@comodo.com>*
8900 * Fix possible deadlock when decoding public keys.
8904 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8906 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8909 * Sanity check record length before skipping explicit IV in DTLS
8912 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8913 fuzzing as a service testing platform.
8918 * Initialise tkeylen properly when encrypting CMS messages.
8919 Thanks to Solar Designer of Openwall for reporting this issue.
8923 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8925 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8926 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8927 in CRYPTO_realloc_clean.
8929 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8930 issue and to Adam Langley <agl@chromium.org> for fixing it.
8933 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8935 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8937 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8938 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8939 content decryption and always return the same error. Note: this attack
8940 needs on average 2^20 messages so it only affects automated senders. The
8941 old behaviour can be re-enabled in the CMS code by setting the
8942 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8943 an MMA defence is not necessary.
8944 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8945 this issue. ([CVE-2012-0884])
8949 * Fix CVE-2011-4619: make sure we really are receiving a
8950 client hello before rejecting multiple SGC restarts. Thanks to
8951 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8955 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8957 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8958 Thanks to Antonio Martin, Enterprise Secure Access Research and
8959 Development, Cisco Systems, Inc. for discovering this bug and
8960 preparing a fix. ([CVE-2012-0050])
8964 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8966 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8967 of the Vaudenay padding oracle attack on CBC mode encryption
8968 which enables an efficient plaintext recovery attack against
8969 the OpenSSL implementation of DTLS. Their attack exploits timing
8970 differences arising during decryption processing. A research
8971 paper describing this attack can be found at:
8972 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8973 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8974 Security Group at Royal Holloway, University of London
8975 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8976 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8977 for preparing the fix. ([CVE-2011-4108])
8979 *Robin Seggelmann, Michael Tuexen*
8981 * Clear bytes used for block padding of SSL 3.0 records.
8984 *Adam Langley (Google)*
8986 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8987 Kadianakis <desnacked@gmail.com> for discovering this issue and
8988 Adam Langley for preparing the fix. ([CVE-2011-4619])
8990 *Adam Langley (Google)*
8992 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8994 *Andrey Kulikov <amdeich@gmail.com>*
8996 * Prevent malformed RFC3779 data triggering an assertion failure.
8997 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8998 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9000 *Rob Austein <sra@hactrn.net>*
9002 * Improved PRNG seeding for VOS.
9004 *Paul Green <Paul.Green@stratus.com>*
9006 * Fix ssl_ciph.c set-up race.
9008 *Adam Langley (Google)*
9010 * Fix spurious failures in ecdsatest.c.
9012 *Emilia Käsper (Google)*
9014 * Fix the BIO_f_buffer() implementation (which was mixing different
9015 interpretations of the `..._len` fields).
9017 *Adam Langley (Google)*
9019 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9020 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9021 threads won't reuse the same blinding coefficients.
9023 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9024 lock to call BN_BLINDING_invert_ex, and avoids one use of
9025 BN_BLINDING_update for each BN_BLINDING structure (previously,
9026 the last update always remained unused).
9028 *Emilia Käsper (Google)*
9030 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9032 *Bob Buckholz (Google)*
9034 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9036 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9037 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9039 *Kaspar Brand <ossl@velox.ch>*
9041 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9042 for multi-threaded use of ECDH. ([CVE-2011-3210])
9044 *Adam Langley (Google)*
9046 * Fix x509_name_ex_d2i memory leak on bad inputs.
9050 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9051 signature public key algorithm by using OID xref utilities instead.
9052 Before this you could only use some ECC ciphersuites with SHA1 only.
9056 * Add protection against ECDSA timing attacks as mentioned in the paper
9057 by Billy Bob Brumley and Nicola Tuveri, see:
9058 <http://eprint.iacr.org/2011/232.pdf>
9060 *Billy Bob Brumley and Nicola Tuveri*
9062 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9064 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9066 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9068 * Fix bug in string printing code: if *any* escaping is enabled we must
9069 escape the escape character (backslash) or the resulting string is
9074 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9076 * Disable code workaround for ancient and obsolete Netscape browsers
9077 and servers: an attacker can use it in a ciphersuite downgrade attack.
9078 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9082 * Fixed J-PAKE implementation error, originally discovered by
9083 Sebastien Martini, further info and confirmation from Stefan
9084 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9088 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9090 * Fix extension code to avoid race conditions which can result in a buffer
9091 overrun vulnerability: resumed sessions must not be modified as they can
9092 be shared by multiple threads. CVE-2010-3864
9096 * Fix WIN32 build system to correctly link an ENGINE directory into
9101 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9103 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9106 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9108 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9110 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9111 context. The operation can be customised via the ctrl mechanism in
9112 case ENGINEs want to include additional functionality.
9116 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9120 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9121 output hashes compatible with older versions of OpenSSL.
9123 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9125 * Fix compression algorithm handling: if resuming a session use the
9126 compression algorithm of the resumed session instead of determining
9127 it from client hello again. Don't allow server to change algorithm.
9131 * Add load_crls() function to commands tidying load_certs() too. Add option
9132 to verify utility to allow additional CRLs to be included.
9136 * Update OCSP request code to permit adding custom headers to the request:
9137 some responders need this.
9141 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9144 *Julia Lawall <julia@diku.dk>*
9146 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9147 needlessly dereferenced structures, used obsolete functions and
9148 didn't handle all updated verify codes correctly.
9152 * Disable MD2 in the default configuration.
9156 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9157 indicate the initial BIO being pushed or popped. This makes it possible
9158 to determine whether the BIO is the one explicitly called or as a result
9159 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9160 it handles reference counts correctly and doesn't zero out the I/O bio
9161 when it is not being explicitly popped. WARNING: applications which
9162 included workarounds for the old buggy behaviour will need to be modified
9163 or they could free up already freed BIOs.
9167 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9168 renaming to all platforms (within the 0.9.8 branch, this was
9169 done conditionally on Netware platforms to avoid a name clash).
9171 *Guenter <lists@gknw.net>*
9173 * Add ECDHE and PSK support to DTLS.
9175 *Michael Tuexen <tuexen@fh-muenster.de>*
9177 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9182 * Add "missing" function EVP_MD_flags() (without this the only way to
9183 retrieve a digest flags is by accessing the structure directly. Update
9184 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9185 or cipher is registered as in the "from" argument. Print out all
9186 registered digests in the dgst usage message instead of manually
9187 attempting to work them out.
9191 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9192 this allows the use of compression and extensions. Change default cipher
9193 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9194 by default unless an application cipher string requests it.
9198 * Alter match criteria in PKCS12_parse(). It used to try to use local
9199 key ids to find matching certificates and keys but some PKCS#12 files
9200 don't follow the (somewhat unwritten) rules and this strategy fails.
9201 Now just gather all certificates together and the first private key
9202 then look for the first certificate that matches the key.
9206 * Support use of registered digest and cipher names for dgst and cipher
9207 commands instead of having to add each one as a special case. So now
9214 openssl dgst -sha256 foo
9216 and this works for ENGINE based algorithms too.
9220 * Update Gost ENGINE to support parameter files.
9222 *Victor B. Wagner <vitus@cryptocom.ru>*
9224 * Support GeneralizedTime in ca utility.
9226 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9228 * Enhance the hash format used for certificate directory links. The new
9229 form uses the canonical encoding (meaning equivalent names will work
9230 even if they aren't identical) and uses SHA1 instead of MD5. This form
9231 is incompatible with the older format and as a result c_rehash should
9232 be used to rebuild symbolic links.
9236 * Make PKCS#8 the default write format for private keys, replacing the
9237 traditional format. This form is standardised, more secure and doesn't
9238 include an implicit MD5 dependency.
9242 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9243 committed to OpenSSL should pass this lot as a minimum.
9247 * Add session ticket override functionality for use by EAP-FAST.
9249 *Jouni Malinen <j@w1.fi>*
9251 * Modify HMAC functions to return a value. Since these can be implemented
9252 in an ENGINE errors can occur.
9256 * Type-checked OBJ_bsearch_ex.
9260 * Type-checked OBJ_bsearch. Also some constification necessitated
9261 by type-checking. Still to come: TXT_DB, bsearch(?),
9262 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9267 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9268 seconds to a tm structure directly, instead of going through OS
9269 specific date routines. This avoids any issues with OS routines such
9270 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9271 and X509_time_adj_ex() to cover the extended range. The existing
9272 X509_time_adj() is still usable and will no longer have any date issues.
9276 * Delta CRL support. New use deltas option which will attempt to locate
9277 and search any appropriate delta CRLs available.
9279 This work was sponsored by Google.
9283 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9284 code and add additional score elements. Validate alternate CRL paths
9285 as part of the CRL checking and indicate a new error "CRL path validation
9286 error" in this case. Applications wanting additional details can use
9287 the verify callback and check the new "parent" field. If this is not
9288 NULL CRL path validation is taking place. Existing applications won't
9289 see this because it requires extended CRL support which is off by
9292 This work was sponsored by Google.
9296 * Support for freshest CRL extension.
9298 This work was sponsored by Google.
9302 * Initial indirect CRL support. Currently only supported in the CRLs
9303 passed directly and not via lookup. Process certificate issuer
9304 CRL entry extension and lookup CRL entries by bother issuer name
9305 and serial number. Check and process CRL issuer entry in IDP extension.
9307 This work was sponsored by Google.
9311 * Add support for distinct certificate and CRL paths. The CRL issuer
9312 certificate is validated separately in this case. Only enabled if
9313 an extended CRL support flag is set: this flag will enable additional
9314 CRL functionality in future.
9316 This work was sponsored by Google.
9320 * Add support for policy mappings extension.
9322 This work was sponsored by Google.
9326 * Fixes to pathlength constraint, self issued certificate handling,
9327 policy processing to align with RFC3280 and PKITS tests.
9329 This work was sponsored by Google.
9333 * Support for name constraints certificate extension. DN, email, DNS
9334 and URI types are currently supported.
9336 This work was sponsored by Google.
9340 * To cater for systems that provide a pointer-based thread ID rather
9341 than numeric, deprecate the current numeric thread ID mechanism and
9342 replace it with a structure and associated callback type. This
9343 mechanism allows a numeric "hash" to be extracted from a thread ID in
9344 either case, and on platforms where pointers are larger than 'long',
9345 mixing is done to help ensure the numeric 'hash' is usable even if it
9346 can't be guaranteed unique. The default mechanism is to use "&errno"
9347 as a pointer-based thread ID to distinguish between threads.
9349 Applications that want to provide their own thread IDs should now use
9350 CRYPTO_THREADID_set_callback() to register a callback that will call
9351 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9353 Note that ERR_remove_state() is now deprecated, because it is tied
9354 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9355 to free the current thread's error state should be replaced by
9356 ERR_remove_thread_state(NULL).
9358 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9359 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9360 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9361 application was previously providing a numeric thread callback that
9362 was inappropriate for distinguishing threads, then uniqueness might
9363 have been obtained with &errno that happened immediately in the
9364 intermediate development versions of OpenSSL; this is no longer the
9365 case, the numeric thread callback will now override the automatic use
9368 *Geoff Thorpe, with help from Bodo Moeller*
9370 * Initial support for different CRL issuing certificates. This covers a
9371 simple case where the self issued certificates in the chain exist and
9372 the real CRL issuer is higher in the existing chain.
9374 This work was sponsored by Google.
9378 * Removed effectively defunct crypto/store from the build.
9382 * Revamp of STACK to provide stronger type-checking. Still to come:
9383 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9384 ASN1_STRING, CONF_VALUE.
9388 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9389 RAM on SSL connections. This option can save about 34k per idle SSL.
9393 * Revamp of LHASH to provide stronger type-checking. Still to come:
9394 STACK, TXT_DB, bsearch, qsort.
9398 * Initial support for Cryptographic Message Syntax (aka CMS) based
9399 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9400 support for data, signedData, compressedData, digestedData and
9401 encryptedData, envelopedData types included. Scripts to check against
9402 RFC4134 examples draft and interop and consistency checks of many
9403 content types and variants.
9407 * Add options to enc utility to support use of zlib compression BIO.
9411 * Extend mk1mf to support importing of options and assembly language
9412 files from Configure script, currently only included in VC-WIN32.
9413 The assembly language rules can now optionally generate the source
9414 files from the associated perl scripts.
9418 * Implement remaining functionality needed to support GOST ciphersuites.
9419 Interop testing has been performed using CryptoPro implementations.
9421 *Victor B. Wagner <vitus@cryptocom.ru>*
9423 * s390x assembler pack.
9427 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9432 * Implement Opaque PRF Input TLS extension as specified in
9433 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9434 official specification yet and no extension type assignment by
9435 IANA exists, this extension (for now) will have to be explicitly
9436 enabled when building OpenSSL by providing the extension number
9437 to use. For example, specify an option
9439 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9441 to the "config" or "Configure" script to enable the extension,
9442 assuming extension number 0x9527 (which is a completely arbitrary
9443 and unofficial assignment based on the MD5 hash of the Internet
9444 Draft). Note that by doing so, you potentially lose
9445 interoperability with other TLS implementations since these might
9446 be using the same extension number for other purposes.
9448 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9449 opaque PRF input value to use in the handshake. This will create
9450 an internal copy of the length-'len' string at 'src', and will
9451 return non-zero for success.
9453 To get more control and flexibility, provide a callback function
9456 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9457 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9461 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9464 Callback function 'cb' will be called in handshakes, and is
9465 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9466 Argument 'arg' is for application purposes (the value as given to
9467 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9468 be provided to the callback function). The callback function
9469 has to return non-zero to report success: usually 1 to use opaque
9470 PRF input just if possible, or 2 to enforce use of the opaque PRF
9471 input. In the latter case, the library will abort the handshake
9472 if opaque PRF input is not successfully negotiated.
9474 Arguments 'peerinput' and 'len' given to the callback function
9475 will always be NULL and 0 in the case of a client. A server will
9476 see the client's opaque PRF input through these variables if
9477 available (NULL and 0 otherwise). Note that if the server
9478 provides an opaque PRF input, the length must be the same as the
9479 length of the client's opaque PRF input.
9481 Note that the callback function will only be called when creating
9482 a new session (session resumption can resume whatever was
9483 previously negotiated), and will not be called in SSL 2.0
9484 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9485 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9486 for applications that need to enforce opaque PRF input.
9490 * Update ssl code to support digests other than SHA1+MD5 for handshake
9493 *Victor B. Wagner <vitus@cryptocom.ru>*
9495 * Add RFC4507 support to OpenSSL. This includes the corrections in
9496 RFC4507bis. The encrypted ticket format is an encrypted encoded
9497 SSL_SESSION structure, that way new session features are automatically
9500 If a client application caches session in an SSL_SESSION structure
9501 support is transparent because tickets are now stored in the encoded
9504 The SSL_CTX structure automatically generates keys for ticket
9505 protection in servers so again support should be possible
9506 with no application modification.
9508 If a client or server wishes to disable RFC4507 support then the option
9509 SSL_OP_NO_TICKET can be set.
9511 Add a TLS extension debugging callback to allow the contents of any client
9512 or server extensions to be examined.
9514 This work was sponsored by Google.
9518 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9519 OpenSSL should now compile cleanly on gcc 4.2
9521 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9523 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9524 support including streaming MAC support: this is required for GOST
9525 ciphersuite support.
9527 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9529 * Add option -stream to use PKCS#7 streaming in smime utility. New
9530 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9531 to output in BER and PEM format.
9535 * Experimental support for use of HMAC via EVP_PKEY interface. This
9536 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9537 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9538 ENGINE support for HMAC keys which are unextractable. New -mac and
9539 -macopt options to dgst utility.
9543 * New option -sigopt to dgst utility. Update dgst to use
9544 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9545 alternative signing parameters such as X9.31 or PSS in the dgst
9550 * Change ssl_cipher_apply_rule(), the internal function that does
9551 the work each time a ciphersuite string requests enabling
9552 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9553 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9554 the order of disabled ciphersuites such that those ciphersuites
9555 that most recently went from enabled to disabled not only stay
9556 in order with respect to each other, but also have higher priority
9557 than other disabled ciphersuites the next time ciphersuites are
9560 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9561 the same ciphersuites as with "HIGH" alone, but in a specific
9562 order where the PSK ciphersuites come first (since they are the
9563 most recently disabled ciphersuites when "HIGH" is parsed).
9565 Also, change ssl_create_cipher_list() (using this new
9566 functionality) such that between otherwise identical
9567 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9572 * Change ssl_create_cipher_list() so that it automatically
9573 arranges the ciphersuites in reasonable order before starting
9574 to process the rule string. Thus, the definition for "DEFAULT"
9575 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9576 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9577 This makes it much easier to arrive at a reasonable default order
9578 in applications for which anonymous ciphers are OK (meaning
9579 that you can't actually use DEFAULT).
9581 *Bodo Moeller; suggested by Victor Duchovni*
9583 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9584 processing) into multiple integers instead of setting
9585 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9586 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9587 (These masks as well as the individual bit definitions are hidden
9588 away into the non-exported interface ssl/ssl_locl.h, so this
9589 change to the definition of the SSL_CIPHER structure shouldn't
9590 affect applications.) This give us more bits for each of these
9591 categories, so there is no longer a need to coagulate AES128 and
9592 AES256 into a single algorithm bit, and to coagulate Camellia128
9593 and Camellia256 into a single algorithm bit, which has led to all
9596 Thus, among other things, the kludge introduced in 0.9.7m and
9597 0.9.8e for masking out AES256 independently of AES128 or masking
9598 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9600 With the change, we also introduce new ciphersuite aliases that
9601 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9606 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9607 Use the leftmost N bytes of the signature input if the input is
9608 larger than the prime q (with N being the size in bytes of q).
9612 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9613 it yet and it is largely untested.
9617 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9621 * Initial incomplete changes to avoid need for function casts in OpenSSL
9622 some compilers (gcc 4.2 and later) reject their use. Safestack is
9623 reimplemented. Update ASN1 to avoid use of legacy functions.
9627 * Win32/64 targets are linked with Winsock2.
9631 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9632 to external functions. This can be used to increase CRL handling
9633 efficiency especially when CRLs are very large by (for example) storing
9634 the CRL revoked certificates in a database.
9638 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9639 new CRLs added to a directory can be used. New command line option
9640 -verify_return_error to s_client and s_server. This causes real errors
9641 to be returned by the verify callback instead of carrying on no matter
9642 what. This reflects the way a "real world" verify callback would behave.
9646 * GOST engine, supporting several GOST algorithms and public key formats.
9647 Kindly donated by Cryptocom.
9651 * Partial support for Issuing Distribution Point CRL extension. CRLs
9652 partitioned by DP are handled but no indirect CRL or reason partitioning
9653 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9654 selected via a scoring technique which handles IDP and AKID in CRLs.
9658 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9659 will ultimately be used for all verify operations: this will remove the
9660 X509_STORE dependency on certificate verification and allow alternative
9661 lookup methods. X509_STORE based implementations of these two callbacks.
9665 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9666 Modify get_crl() to find a valid (unexpired) CRL if possible.
9670 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9671 this would be called X509_CRL_cmp() but that name is already used by
9672 a function that just compares CRL issuer names. Cache several CRL
9673 extensions in X509_CRL structure and cache CRLDP in X509.
9677 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9678 this maps equivalent X509_NAME structures into a consistent structure.
9679 Name comparison can then be performed rapidly using memcmp().
9683 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9688 * Allow digests to supply their own micalg string for S/MIME type using
9689 the ctrl EVP_MD_CTRL_MICALG.
9693 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9694 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9695 ctrl. It can then customise the structure before and/or after signing
9700 * New function OBJ_add_sigid() to allow application defined signature OIDs
9701 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9702 to free up any added signature OIDs.
9706 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9707 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9708 digest and cipher tables. New options added to openssl utility:
9709 list-message-digest-algorithms and list-cipher-algorithms.
9713 * Change the array representation of binary polynomials: the list
9714 of degrees of non-zero coefficients is now terminated with -1.
9715 Previously it was terminated with 0, which was also part of the
9716 value; thus, the array representation was not applicable to
9717 polynomials where t^0 has coefficient zero. This change makes
9718 the array representation useful in a more general context.
9722 * Various modifications and fixes to SSL/TLS cipher string
9723 handling. For ECC, the code now distinguishes between fixed ECDH
9724 with RSA certificates on the one hand and with ECDSA certificates
9725 on the other hand, since these are separate ciphersuites. The
9726 unused code for Fortezza ciphersuites has been removed.
9728 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9729 (not "ECDHE"). For consistency with the code for DH
9730 certificates, use of ECDH certificates is now considered ECDH
9731 authentication, not RSA or ECDSA authentication (the latter is
9732 merely the CA's signing algorithm and not actively used in the
9735 The temporary ciphersuite alias "ECCdraft" is no longer
9736 available, and ECC ciphersuites are no longer excluded from "ALL"
9737 and "DEFAULT". The following aliases now exist for RFC 4492
9738 ciphersuites, most of these by analogy with the DH case:
9740 kECDHr - ECDH cert, signed with RSA
9741 kECDHe - ECDH cert, signed with ECDSA
9742 kECDH - ECDH cert (signed with either RSA or ECDSA)
9743 kEECDH - ephemeral ECDH
9744 ECDH - ECDH cert or ephemeral ECDH
9750 AECDH - anonymous ECDH
9751 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9755 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9756 Use correct micalg parameters depending on digest(s) in signed message.
9760 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9761 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9765 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9766 an engine to register a method. Add ENGINE lookups for methods and
9767 functional reference processing.
9771 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9772 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9777 * New -resign option to smime utility. This adds one or more signers
9778 to an existing PKCS#7 signedData structure. Also -md option to use an
9779 alternative message digest algorithm for signing.
9783 * Tidy up PKCS#7 routines and add new functions to make it easier to
9784 create PKCS7 structures containing multiple signers. Update smime
9785 application to support multiple signers.
9789 * New -macalg option to pkcs12 utility to allow setting of an alternative
9794 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9795 Reorganize PBE internals to lookup from a static table using NIDs,
9796 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9797 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9798 PRF which will be automatically used with PBES2.
9802 * Replace the algorithm specific calls to generate keys in "req" with the
9807 * Update PKCS#7 enveloped data routines to use new API. This is now
9808 supported by any public key method supporting the encrypt operation. A
9809 ctrl is added to allow the public key algorithm to examine or modify
9810 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9815 * Add a ctrl to asn1 method to allow a public key algorithm to express
9816 a default digest type to use. In most cases this will be SHA1 but some
9817 algorithms (such as GOST) need to specify an alternative digest. The
9818 return value indicates how strong the preference is 1 means optional and
9819 2 is mandatory (that is it is the only supported type). Modify
9820 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9821 use the default md. Update openssl utilities to use the default digest
9822 type for signing if it is not explicitly indicated.
9826 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9827 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9828 signing method from the key type. This effectively removes the link
9829 between digests and public key types.
9833 * Add an OID cross reference table and utility functions. Its purpose is to
9834 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9835 rsaEncryption. This will allow some of the algorithm specific hackery
9836 needed to use the correct OID to be removed.
9840 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9841 structures for PKCS7_sign(). They are now set up by the relevant public
9846 * Add provisional EC pkey method with support for ECDSA and ECDH.
9850 * Add support for key derivation (agreement) in the API, DH method and
9855 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9856 public and private key formats. As a side effect these add additional
9857 command line functionality not previously available: DSA signatures can be
9858 generated and verified using pkeyutl and DH key support and generation in
9865 *Oliver Tappe <zooey@hirschkaefer.de>*
9867 * New make target "install_html_docs" installs HTML renditions of the
9870 *Oliver Tappe <zooey@hirschkaefer.de>*
9872 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9873 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9874 support key and parameter generation and add initial key generation
9875 functionality for RSA.
9879 * Add functions for main EVP_PKEY_method operations. The undocumented
9880 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9881 `EVP_PKEY_{encrypt,decrypt}_old`.
9885 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9886 key API, doesn't do much yet.
9890 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9891 public key algorithms. New option to openssl utility:
9892 "list-public-key-algorithms" to print out info.
9896 * Implement the Supported Elliptic Curves Extension for
9897 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9901 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9902 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9906 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9907 utilities such as rsa, dsa, dsaparam etc except they process any key
9912 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9913 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9914 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9919 * Initial support for pluggable public key ASN1.
9920 De-spaghettify the public key ASN1 handling. Move public and private
9921 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9922 algorithm specific handling to a single module within the relevant
9923 algorithm directory. Add functions to allow (near) opaque processing
9924 of public and private key structures.
9928 * Implement the Supported Point Formats Extension for
9929 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9933 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9934 for the psk identity [hint] and the psk callback functions to the
9935 SSL_SESSION, SSL and SSL_CTX structure.
9938 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9942 SSL_CTX_use_psk_identity_hint
9943 SSL_get_psk_identity_hint
9944 SSL_get_psk_identity
9945 SSL_use_psk_identity_hint
9947 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9949 * Add RFC 3161 compliant time stamp request creation, response generation
9950 and response verification functionality.
9952 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9954 * Add initial support for TLS extensions, specifically for the server_name
9955 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9956 have new members for a hostname. The SSL data structure has an
9957 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9958 stored in that context to allow for session resumption, even after the
9959 SSL has been switched to a new SSL_CTX in reaction to a client's
9960 server_name extension.
9962 New functions (subject to change):
9964 SSL_get_servername()
9965 SSL_get_servername_type()
9968 New CTRL codes and macros (subject to change):
9970 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9971 - SSL_CTX_set_tlsext_servername_callback()
9972 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9973 - SSL_CTX_set_tlsext_servername_arg()
9974 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9976 openssl s_client has a new '-servername ...' option.
9978 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9979 '-key2 ...', '-servername_fatal' (subject to change). This allows
9980 testing the HostName extension for a specific single hostname ('-cert'
9981 and '-key' remain fallbacks for handshakes without HostName
9982 negotiation). If the unrecognized_name alert has to be sent, this by
9983 default is a warning; it becomes fatal with the '-servername_fatal'
9986 *Peter Sylvester, Remy Allais, Christophe Renou*
9988 * Whirlpool hash implementation is added.
9992 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9993 bn(64,32). Because of instruction set limitations it doesn't have
9994 any negative impact on performance. This was done mostly in order
9995 to make it possible to share assembler modules, such as bn_mul_mont
9996 implementations, between 32- and 64-bit builds without hassle.
10000 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10001 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10006 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10007 dedicated Montgomery multiplication procedure, is introduced.
10008 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10009 "64-bit" performance on certain 32-bit targets.
10013 * New option SSL_OP_NO_COMP to disable use of compression selectively
10014 in SSL structures. New SSL ctrl to set maximum send fragment size.
10015 Save memory by setting the I/O buffer sizes dynamically instead of
10016 using the maximum available value.
10020 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10021 in addition to the text details.
10025 * Very, very preliminary EXPERIMENTAL support for printing of general
10026 ASN1 structures. This currently produces rather ugly output and doesn't
10027 handle several customised structures at all.
10031 * Integrated support for PVK file format and some related formats such
10032 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10033 these in the 'rsa' and 'dsa' utilities.
10037 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10041 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10042 place for the (very old) "NETSCAPE" format certificates which are now
10043 handled using new ASN1 code equivalents.
10047 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10048 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10049 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10053 * Modify CRL distribution points extension code to print out previously
10054 unsupported fields. Enhance extension setting code to allow setting of
10059 * Add print and set support for Issuing Distribution Point CRL extension.
10063 * Change 'Configure' script to enable Camellia by default.
10070 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10072 * When rejecting SSL/TLS records due to an incorrect version number, never
10073 update s->server with a new major version number. As of
10074 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10075 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10076 the previous behavior could result in a read attempt at NULL when
10077 receiving specific incorrect SSL/TLS records once record payload
10078 protection is active. ([CVE-2010-0740])
10080 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10082 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10083 could be crashed if the relevant tables were not present (e.g. chrooted).
10085 *Tomas Hoger <thoger@redhat.com>*
10087 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10089 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10091 *Martin Olsson, Neel Mehta*
10093 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10094 accommodate for stack sorting, always a write lock!).
10098 * On some versions of WIN32 Heap32Next is very slow. This can cause
10099 excessive delays in the RAND_poll(): over a minute. As a workaround
10100 include a time check in the inner Heap32Next loop too.
10104 * The code that handled flushing of data in SSL/TLS originally used the
10105 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10106 the problem outlined in PR#1949. The fix suggested there however can
10107 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10108 of Apache). So instead simplify the code to flush unconditionally.
10109 This should be fine since flushing with no data to flush is a no op.
10113 * Handle TLS versions 2.0 and later properly and correctly use the
10114 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10115 off ancient servers have a habit of sticking around for a while...
10119 * Modify compression code so it frees up structures without using the
10120 ex_data callbacks. This works around a problem where some applications
10121 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10122 restarting) then use compression (e.g. SSL with compression) later.
10123 This results in significant per-connection memory leaks and
10124 has caused some security issues including CVE-2008-1678 and
10129 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10130 change when encrypting or decrypting.
10134 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10135 connect and renegotiate with servers which do not support RI.
10136 Until RI is more widely deployed this option is enabled by default.
10140 * Add "missing" ssl ctrls to clear options and mode.
10144 * If client attempts to renegotiate and doesn't support RI respond with
10145 a no_renegotiation alert as required by RFC5746. Some renegotiating
10146 TLS clients will continue a connection gracefully when they receive
10147 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10148 waiting for a server hello which it will never receive. Now we treat a
10149 received no_renegotiation alert as a fatal error. This is because
10150 applications requesting a renegotiation might well expect it to succeed
10151 and would have no code in place to handle the server denying it so the
10152 only safe thing to do is to terminate the connection.
10156 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10157 peer supports secure renegotiation and 0 otherwise. Print out peer
10158 renegotiation support in s_client/s_server.
10162 * Replace the highly broken and deprecated SPKAC certification method with
10163 the updated NID creation version. This should correctly handle UTF8.
10167 * Implement RFC5746. Re-enable renegotiation but require the extension
10168 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10169 turns out to be a bad idea. It has been replaced by
10170 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10171 SSL_CTX_set_options(). This is really not recommended unless you
10172 know what you are doing.
10174 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10176 * Fixes to stateless session resumption handling. Use initial_ctx when
10177 issuing and attempting to decrypt tickets in case it has changed during
10178 servername handling. Use a non-zero length session ID when attempting
10179 stateless session resumption: this makes it possible to determine if
10180 a resumption has occurred immediately after receiving server hello
10181 (several places in OpenSSL subtly assume this) instead of later in
10186 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10187 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10188 fixes for a few places where the return code is not checked
10191 *Julia Lawall <julia@diku.dk>*
10193 * Add --strict-warnings option to Configure script to include devteam
10194 warnings in other configurations.
10198 * Add support for --libdir option and LIBDIR variable in makefiles. This
10199 makes it possible to install openssl libraries in locations which
10200 have names other than "lib", for example "/usr/lib64" which some
10203 *Steve Henson, based on patch from Jeremy Utley*
10205 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10206 X690 8.9.12 and can produce some misleading textual output of OIDs.
10208 *Steve Henson, reported by Dan Kaminsky*
10210 * Delete MD2 from algorithm tables. This follows the recommendation in
10211 several standards that it is not used in new applications due to
10212 several cryptographic weaknesses. For binary compatibility reasons
10213 the MD2 API is still compiled in by default.
10217 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10222 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10223 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10226 *Guenter <lists@gknw.net>*
10228 * Fix the server certificate chain building code to use X509_verify_cert(),
10229 it used to have an ad-hoc builder which was unable to cope with anything
10230 other than a simple chain.
10232 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10234 * Don't check self signed certificate signatures in X509_verify_cert()
10235 by default (a flag can override this): it just wastes time without
10236 adding any security. As a useful side effect self signed root CAs
10237 with non-FIPS digests are now usable in FIPS mode.
10241 * In dtls1_process_out_of_seq_message() the check if the current message
10242 is already buffered was missing. For every new message was memory
10243 allocated, allowing an attacker to perform an denial of service attack
10244 with sending out of seq handshake messages until there is no memory
10245 left. Additionally every future message was buffered, even if the
10246 sequence number made no sense and would be part of another handshake.
10247 So only messages with sequence numbers less than 10 in advance will be
10248 buffered. ([CVE-2009-1378])
10250 *Robin Seggelmann, discovered by Daniel Mentz*
10252 * Records are buffered if they arrive with a future epoch to be
10253 processed after finishing the corresponding handshake. There is
10254 currently no limitation to this buffer allowing an attacker to perform
10255 a DOS attack with sending records with future epochs until there is no
10256 memory left. This patch adds the pqueue_size() function to determine
10257 the size of a buffer and limits the record buffer to 100 entries.
10260 *Robin Seggelmann, discovered by Daniel Mentz*
10262 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10263 parent structure is freed. ([CVE-2009-1379])
10267 * Handle non-blocking I/O properly in SSL_shutdown() call.
10269 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10271 * Add `2.5.4.*` OIDs
10273 *Ilya O. <vrghost@gmail.com>*
10275 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10277 * Disable renegotiation completely - this fixes a severe security
10278 problem ([CVE-2009-3555]) at the cost of breaking all
10279 renegotiation. Renegotiation can be re-enabled by setting
10280 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10281 run-time. This is really not recommended unless you know what
10286 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10288 * Don't set val to NULL when freeing up structures, it is freed up by
10289 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10290 zeroing past the valid field. ([CVE-2009-0789])
10292 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10294 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10295 checked correctly. This would allow some invalid signed attributes to
10296 appear to verify correctly. ([CVE-2009-0591])
10298 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10300 * Reject UniversalString and BMPString types with invalid lengths. This
10301 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10302 a legal length. ([CVE-2009-0590])
10306 * Set S/MIME signing as the default purpose rather than setting it
10307 unconditionally. This allows applications to override it at the store
10312 * Permit restricted recursion of ASN1 strings. This is needed in practice
10313 to handle some structures.
10317 * Improve efficiency of mem_gets: don't search whole buffer each time
10320 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10322 * New -hex option for openssl rand.
10326 * Print out UTF8String and NumericString when parsing ASN1.
10330 * Support NumericString type for name components.
10334 * Allow CC in the environment to override the automatically chosen
10335 compiler. Note that nothing is done to ensure flags work with the
10340 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10342 * Properly check EVP_VerifyFinal() and similar return values
10345 *Ben Laurie, Bodo Moeller, Google Security Team*
10347 * Enable TLS extensions by default.
10351 * Allow the CHIL engine to be loaded, whether the application is
10352 multithreaded or not. (This does not release the developer from the
10353 obligation to set up the dynamic locking callbacks.)
10355 *Sander Temme <sander@temme.net>*
10357 * Use correct exit code if there is an error in dgst command.
10359 *Steve Henson; problem pointed out by Roland Dirlewanger*
10361 * Tweak Configure so that you need to say "experimental-jpake" to enable
10362 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10366 * Add experimental JPAKE support, including demo authentication in
10367 s_client and s_server.
10371 * Set the comparison function in v3_addr_canonize().
10373 *Rob Austein <sra@hactrn.net>*
10375 * Add support for XMPP STARTTLS in s_client.
10377 *Philip Paeps <philip@freebsd.org>*
10379 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10380 to ensure that even with this option, only ciphersuites in the
10381 server's preference list will be accepted. (Note that the option
10382 applies only when resuming a session, so the earlier behavior was
10383 just about the algorithm choice for symmetric cryptography.)
10387 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10389 * Fix NULL pointer dereference if a DTLS server received
10390 ChangeCipherSpec as first record ([CVE-2009-1386]).
10394 * Fix a state transition in s3_srvr.c and d1_srvr.c
10395 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10397 *Nagendra Modadugu*
10399 * The fix in 0.9.8c that supposedly got rid of unsafe
10400 double-checked locking was incomplete for RSA blinding,
10401 addressing just one layer of what turns out to have been
10402 doubly unsafe triple-checked locking.
10404 So now fix this for real by retiring the MONT_HELPER macro
10405 in crypto/rsa/rsa_eay.c.
10407 *Bodo Moeller; problem pointed out by Marius Schilder*
10409 * Various precautionary measures:
10411 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10413 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10414 (NB: This would require knowledge of the secret session ticket key
10415 to exploit, in which case you'd be SOL either way.)
10417 - Change bn_nist.c so that it will properly handle input BIGNUMs
10418 outside the expected range.
10420 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10423 *Neel Mehta, Bodo Moeller*
10425 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10426 the load fails. Useful for distros.
10428 *Ben Laurie and the FreeBSD team*
10430 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10434 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10438 * Expand ENGINE to support engine supplied SSL client certificate functions.
10440 This work was sponsored by Logica.
10444 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10445 keystores. Support for SSL/TLS client authentication too.
10446 Not compiled unless enable-capieng specified to Configure.
10448 This work was sponsored by Logica.
10452 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10453 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10454 attribute creation routines such as certificate requests and PKCS#12
10459 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10461 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10462 handshake which could lead to a client crash as found using the
10463 Codenomicon TLS test suite ([CVE-2008-1672])
10465 *Steve Henson, Mark Cox*
10467 * Fix double free in TLS server name extensions which could lead to
10468 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10472 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10474 Clear the error queue to ensure that error entries left from
10475 older function calls do not interfere with the correct operation.
10477 *Lutz Jaenicke, Erik de Castro Lopo*
10479 * Remove root CA certificates of commercial CAs:
10481 The OpenSSL project does not recommend any specific CA and does not
10482 have any policy with respect to including or excluding any CA.
10483 Therefore, it does not make any sense to ship an arbitrary selection
10484 of root CA certificates with the OpenSSL software.
10488 * RSA OAEP patches to fix two separate invalid memory reads.
10489 The first one involves inputs when 'lzero' is greater than
10490 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10491 before the beginning of from). The second one involves inputs where
10492 the 'db' section contains nothing but zeroes (there is a one-byte
10493 invalid read after the end of 'db').
10495 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10497 * Partial backport from 0.9.9-dev:
10499 Introduce bn_mul_mont (dedicated Montgomery multiplication
10500 procedure) as a candidate for BIGNUM assembler implementation.
10501 While 0.9.9-dev uses assembler for various architectures, only
10502 x86_64 is available by default here in the 0.9.8 branch, and
10503 32-bit x86 is available through a compile-time setting.
10505 To try the 32-bit x86 assembler implementation, use Configure
10506 option "enable-montasm" (which exists only for this backport).
10508 As "enable-montasm" for 32-bit x86 disclaims code stability
10509 anyway, in this constellation we activate additional code
10510 backported from 0.9.9-dev for further performance improvements,
10511 namely BN_from_montgomery_word. (To enable this otherwise,
10512 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10514 *Andy Polyakov (backport partially by Bodo Moeller)*
10516 * Add TLS session ticket callback. This allows an application to set
10517 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10518 values. This is useful for key rollover for example where several key
10519 sets may exist with different names.
10523 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10524 This was broken until now in 0.9.8 releases, such that the only way
10525 a registered ENGINE could be used (assuming it initialises
10526 successfully on the host) was to explicitly set it as the default
10527 for the relevant algorithms. This is in contradiction with 0.9.7
10528 behaviour and the documentation. With this fix, when an ENGINE is
10529 registered into a given algorithm's table of implementations, the
10530 'uptodate' flag is reset so that auto-discovery will be used next
10531 time a new context for that algorithm attempts to select an
10534 *Ian Lister (tweaked by Geoff Thorpe)*
10536 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10537 implementation in the following ways:
10539 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10542 Lack of BER streaming support means one pass streaming processing is
10543 only supported if data is detached: setting the streaming flag is
10544 ignored for embedded content.
10546 CMS support is disabled by default and must be explicitly enabled
10547 with the enable-cms configuration option.
10551 * Update the GMP engine glue to do direct copies between BIGNUM and
10552 mpz_t when openssl and GMP use the same limb size. Otherwise the
10553 existing "conversion via a text string export" trick is still used.
10555 *Paul Sheer <paulsheer@gmail.com>*
10557 * Zlib compression BIO. This is a filter BIO which compressed and
10558 uncompresses any data passed through it.
10562 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10563 RFC3394 compatible AES key wrapping.
10567 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10568 sets string data without copying. X509_ALGOR_set0() and
10569 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10570 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10571 from an X509_ATTRIBUTE structure optionally checking it occurs only
10572 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10577 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10578 to get the expected BN_FLG_CONSTTIME behavior.
10580 *Bodo Moeller (Google)*
10584 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10585 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10586 - added some more tests to do_tests.pl
10587 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10588 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10589 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10590 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10591 - various changes to netware.pl to enable gcc-cross builds on Win32
10593 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10594 - various changes to fix missing prototype warnings
10595 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10596 - added AES, WHIRLPOOL and CPUID assembler code to build files
10597 - added missing AES assembler make rules to mk1mf.pl
10598 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10600 *Guenter Knauf <eflash@gmx.net>*
10602 * Implement certificate status request TLS extension defined in RFC3546.
10603 A client can set the appropriate parameters and receive the encoded
10604 OCSP response via a callback. A server can query the supplied parameters
10605 and set the encoded OCSP response in the callback. Add simplified examples
10606 to s_client and s_server.
10610 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10612 * Fix various bugs:
10613 + Binary incompatibility of ssl_ctx_st structure
10614 + DTLS interoperation with non-compliant servers
10615 + Don't call get_session_cb() without proposed session
10616 + Fix ia64 assembler code
10618 *Andy Polyakov, Steve Henson*
10620 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10622 * DTLS Handshake overhaul. There were longstanding issues with
10623 OpenSSL DTLS implementation, which were making it impossible for
10624 RFC 4347 compliant client to communicate with OpenSSL server.
10625 Unfortunately just fixing these incompatibilities would "cut off"
10626 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10627 server keeps tolerating non RFC compliant syntax. The opposite is
10628 not true, 0.9.8f client can not communicate with earlier server.
10629 This update even addresses CVE-2007-4995.
10633 * Changes to avoid need for function casts in OpenSSL: some compilers
10634 (gcc 4.2 and later) reject their use.
10635 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10638 * Add RFC4507 support to OpenSSL. This includes the corrections in
10639 RFC4507bis. The encrypted ticket format is an encrypted encoded
10640 SSL_SESSION structure, that way new session features are automatically
10643 If a client application caches session in an SSL_SESSION structure
10644 support is transparent because tickets are now stored in the encoded
10647 The SSL_CTX structure automatically generates keys for ticket
10648 protection in servers so again support should be possible
10649 with no application modification.
10651 If a client or server wishes to disable RFC4507 support then the option
10652 SSL_OP_NO_TICKET can be set.
10654 Add a TLS extension debugging callback to allow the contents of any client
10655 or server extensions to be examined.
10657 This work was sponsored by Google.
10661 * Add initial support for TLS extensions, specifically for the server_name
10662 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10663 have new members for a hostname. The SSL data structure has an
10664 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10665 stored in that context to allow for session resumption, even after the
10666 SSL has been switched to a new SSL_CTX in reaction to a client's
10667 server_name extension.
10669 New functions (subject to change):
10671 SSL_get_servername()
10672 SSL_get_servername_type()
10675 New CTRL codes and macros (subject to change):
10677 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10678 - SSL_CTX_set_tlsext_servername_callback()
10679 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10680 - SSL_CTX_set_tlsext_servername_arg()
10681 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10683 openssl s_client has a new '-servername ...' option.
10685 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10686 '-key2 ...', '-servername_fatal' (subject to change). This allows
10687 testing the HostName extension for a specific single hostname ('-cert'
10688 and '-key' remain fallbacks for handshakes without HostName
10689 negotiation). If the unrecognized_name alert has to be sent, this by
10690 default is a warning; it becomes fatal with the '-servername_fatal'
10693 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10695 * Add AES and SSE2 assembly language support to VC++ build.
10699 * Mitigate attack on final subtraction in Montgomery reduction.
10703 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10704 (which previously caused an internal error).
10708 * Squeeze another 10% out of IGE mode when in != out.
10712 * AES IGE mode speedup.
10714 *Dean Gaudet (Google)*
10716 * Add the Korean symmetric 128-bit cipher SEED (see
10717 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10718 add SEED ciphersuites from RFC 4162:
10720 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10721 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10722 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10723 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10725 To minimize changes between patchlevels in the OpenSSL 0.9.8
10726 series, SEED remains excluded from compilation unless OpenSSL
10727 is configured with 'enable-seed'.
10729 *KISA, Bodo Moeller*
10731 * Mitigate branch prediction attacks, which can be practical if a
10732 single processor is shared, allowing a spy process to extract
10733 information. For detailed background information, see
10734 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10735 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10736 and Necessary Software Countermeasures"). The core of the change
10737 are new versions BN_div_no_branch() and
10738 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10739 respectively, which are slower, but avoid the security-relevant
10740 conditional branches. These are automatically called by BN_div()
10741 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10742 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10743 remove a conditional branch.
10745 BN_FLG_CONSTTIME is the new name for the previous
10746 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10747 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10748 in the exponent causes BN_mod_exp_mont() to use the alternative
10749 implementation in BN_mod_exp_mont_consttime().) The old name
10750 remains as a deprecated alias.
10752 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10753 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10754 constant-time implementations for more than just exponentiation.
10755 Here too the old name is kept as a deprecated alias.
10757 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10758 the BN_BLINDING structure gets an independent copy of the
10759 modulus. This means that the previous `BIGNUM *m` argument to
10760 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10761 essentially becomes `const BIGNUM *m`, although we can't actually
10762 change this in the header file before 0.9.9. It allows
10763 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10764 enable BN_FLG_CONSTTIME.
10766 *Matthew D Wood (Intel Corp)*
10768 * In the SSL/TLS server implementation, be strict about session ID
10769 context matching (which matters if an application uses a single
10770 external cache for different purposes). Previously,
10771 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10772 set. This did ensure strict client verification, but meant that,
10773 with applications using a single external cache for quite
10774 different requirements, clients could circumvent ciphersuite
10775 restrictions for a given session ID context by starting a session
10776 in a different context.
10780 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10781 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10782 authentication-only ciphersuites.
10786 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10787 not complete and could lead to a possible single byte overflow
10788 ([CVE-2007-5135]) [Ben Laurie]
10790 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10792 * Since AES128 and AES256 (and similarly Camellia128 and
10793 Camellia256) share a single mask bit in the logic of
10794 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10795 kludge to work properly if AES128 is available and AES256 isn't
10796 (or if Camellia128 is available and Camellia256 isn't).
10800 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10801 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10802 When a point or a seed is encoded in a BIT STRING, we need to
10803 prevent the removal of trailing zero bits to get the proper DER
10804 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10805 of a NamedBitList, for which trailing 0 bits need to be removed.)
10809 * Have SSL/TLS server implementation tolerate "mismatched" record
10810 protocol version while receiving ClientHello even if the
10811 ClientHello is fragmented. (The server can't insist on the
10812 particular protocol version it has chosen before the ServerHello
10813 message has informed the client about his choice.)
10817 * Add RFC 3779 support.
10819 *Rob Austein for ARIN, Ben Laurie*
10821 * Load error codes if they are not already present instead of using a
10822 static variable. This allows them to be cleanly unloaded and reloaded.
10823 Improve header file function name parsing.
10827 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10828 or CAPABILITY handshake as required by RFCs.
10830 *Goetz Babin-Ebell*
10832 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10834 * Introduce limits to prevent malicious keys being able to
10835 cause a denial of service. ([CVE-2006-2940])
10837 *Steve Henson, Bodo Moeller*
10839 * Fix ASN.1 parsing of certain invalid structures that can result
10840 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10842 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10843 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10845 * Fix SSL client code which could crash if connecting to a
10846 malicious SSLv2 server. ([CVE-2006-4343])
10848 *Tavis Ormandy and Will Drewry, Google Security Team*
10850 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10851 match only those. Before that, "AES256-SHA" would be interpreted
10852 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10853 the same strength classification in 0.9.7h) as we currently only
10854 have a single AES bit in the ciphersuite description bitmap.
10855 That change, however, also applied to ciphersuite strings such as
10856 "RC4-MD5" that intentionally matched multiple ciphersuites --
10857 namely, SSL 2.0 ciphersuites in addition to the more common ones
10858 from SSL 3.0/TLS 1.0.
10860 So we change the selection algorithm again: Naming an explicit
10861 ciphersuite selects this one ciphersuite, and any other similar
10862 ciphersuite (same bitmap) from *other* protocol versions.
10863 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10864 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10866 Since SSL 2.0 does not have any ciphersuites for which the
10867 128/256 bit distinction would be relevant, this works for now.
10868 The proper fix will be to use different bits for AES128 and
10869 AES256, which would have avoided the problems from the beginning;
10870 however, bits are scarce, so we can only do this in a new release
10871 (not just a patchlevel) when we can change the SSL_CIPHER
10872 definition to split the single 'unsigned long mask' bitmap into
10873 multiple values to extend the available space.
10877 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10879 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10880 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10882 * Add AES IGE and biIGE modes.
10886 * Change the Unix randomness entropy gathering to use poll() when
10887 possible instead of select(), since the latter has some
10888 undesirable limitations.
10890 *Darryl Miles via Richard Levitte and Bodo Moeller*
10892 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10893 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10894 cannot be implicitly activated as part of, e.g., the "AES" alias.
10895 However, please upgrade to OpenSSL 0.9.9[-dev] for
10896 non-experimental use of the ECC ciphersuites to get TLS extension
10897 support, which is required for curve and point format negotiation
10898 to avoid potential handshake problems.
10902 * Disable rogue ciphersuites:
10904 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10905 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10906 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10908 The latter two were purportedly from
10909 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10912 Also deactivate the remaining ciphersuites from
10913 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10914 unofficial, and the ID has long expired.
10918 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10919 dual-core machines) and other potential thread-safety issues.
10923 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10924 versions), which is now available for royalty-free use
10925 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10926 Also, add Camellia TLS ciphersuites from RFC 4132.
10928 To minimize changes between patchlevels in the OpenSSL 0.9.8
10929 series, Camellia remains excluded from compilation unless OpenSSL
10930 is configured with 'enable-camellia'.
10934 * Disable the padding bug check when compression is in use. The padding
10935 bug check assumes the first packet is of even length, this is not
10936 necessarily true if compression is enabled and can result in false
10937 positives causing handshake failure. The actual bug test is ancient
10938 code so it is hoped that implementations will either have fixed it by
10939 now or any which still have the bug do not support compression.
10943 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10945 * When applying a cipher rule check to see if string match is an explicit
10946 cipher suite and only match that one cipher suite if it is.
10950 * Link in manifests for VC++ if needed.
10952 *Austin Ziegler <halostatue@gmail.com>*
10954 * Update support for ECC-based TLS ciphersuites according to
10955 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10956 TLS extensions, which are supported starting with the 0.9.9
10957 branch, not in the OpenSSL 0.9.8 branch).
10961 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10962 opaque EVP_CIPHER_CTX handling.
10966 * Fixes and enhancements to zlib compression code. We now only use
10967 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10968 to conform with the standards mentioned here:
10969 <http://www.zlib.net/DLL_FAQ.txt>
10970 Static zlib linking now works on Windows and the new --with-zlib-include
10971 --with-zlib-lib options to Configure can be used to supply the location
10972 of the headers and library. Gracefully handle case where zlib library
10977 * Several fixes and enhancements to the OID generation code. The old code
10978 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10979 handle numbers larger than ULONG_MAX, truncated printing and had a
10980 non standard OBJ_obj2txt() behaviour.
10984 * Add support for building of engines under engine/ as shared libraries
10985 under VC++ build system.
10989 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10990 Hopefully, we will not see any false combination of paths any more.
10994 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10996 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10997 (part of SSL_OP_ALL). This option used to disable the
10998 countermeasure against man-in-the-middle protocol-version
10999 rollback in the SSL 2.0 server implementation, which is a bad
11000 idea. ([CVE-2005-2969])
11002 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11003 for Information Security, National Institute of Advanced Industrial
11004 Science and Technology [AIST], Japan)*
11006 * Add two function to clear and return the verify parameter flags.
11010 * Keep cipherlists sorted in the source instead of sorting them at
11011 runtime, thus removing the need for a lock.
11015 * Avoid some small subgroup attacks in Diffie-Hellman.
11017 *Nick Mathewson and Ben Laurie*
11019 * Add functions for well-known primes.
11023 * Extended Windows CE support.
11025 *Satoshi Nakamura and Andy Polyakov*
11027 * Initialize SSL_METHOD structures at compile time instead of during
11028 runtime, thus removing the need for a lock.
11032 * Make PKCS7_decrypt() work even if no certificate is supplied by
11033 attempting to decrypt each encrypted key in turn. Add support to
11038 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11040 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11043 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11047 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11048 key into the same file any more.
11052 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11056 * Add -utf8 command line and config file option to 'ca'.
11058 *Stefan <stf@udoma.org*
11060 * Removed the macro des_crypt(), as it seems to conflict with some
11061 libraries. Use DES_crypt().
11065 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11066 involves renaming the source and generated shared-libs for
11067 both. The engines will accept the corrected or legacy ids
11068 ('ncipher' and '4758_cca' respectively) when binding. NB,
11069 this only applies when building 'shared'.
11071 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11073 * Add attribute functions to EVP_PKEY structure. Modify
11074 PKCS12_create() to recognize a CSP name attribute and
11075 use it. Make -CSP option work again in pkcs12 utility.
11079 * Add new functionality to the bn blinding code:
11080 - automatic re-creation of the BN_BLINDING parameters after
11081 a fixed number of uses (currently 32)
11082 - add new function for parameter creation
11083 - introduce flags to control the update behaviour of the
11084 BN_BLINDING parameters
11085 - hide BN_BLINDING structure
11086 Add a second BN_BLINDING slot to the RSA structure to improve
11087 performance when a single RSA object is shared among several
11092 * Add support for DTLS.
11094 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11096 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11097 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11101 * Remove buggy and incomplete DH cert support from
11102 ssl/ssl_rsa.c and ssl/s3_both.c
11106 * Use SHA-1 instead of MD5 as the default digest algorithm for
11107 the `apps/openssl` commands.
11111 * Compile clean with "-Wall -Wmissing-prototypes
11112 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11113 DEBUG_SAFESTACK must also be set.
11117 * Change ./Configure so that certain algorithms can be disabled by default.
11118 The new counterpiece to "no-xxx" is "enable-xxx".
11120 The patented RC5 and MDC2 algorithms will now be disabled unless
11121 "enable-rc5" and "enable-mdc2", respectively, are specified.
11123 (IDEA remains enabled despite being patented. This is because IDEA
11124 is frequently required for interoperability, and there is no license
11125 fee for non-commercial use. As before, "no-idea" can be used to
11126 avoid this algorithm.)
11130 * Add processing of proxy certificates (see RFC 3820). This work was
11131 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11132 EGEE (Enabling Grids for E-science in Europe).
11136 * RC4 performance overhaul on modern architectures/implementations, such
11137 as Intel P4, IA-64 and AMD64.
11141 * New utility extract-section.pl. This can be used specify an alternative
11142 section number in a pod file instead of having to treat each file as
11143 a separate case in Makefile. This can be done by adding two lines to the
11146 =for comment openssl_section:XXX
11148 The blank line is mandatory.
11152 * New arguments -certform, -keyform and -pass for s_client and s_server
11153 to allow alternative format key and certificate files and passphrase
11158 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11159 update associated structures and add various utility functions.
11161 Add new policy related verify parameters, include policy checking in
11162 standard verify code. Enhance 'smime' application with extra parameters
11163 to support policy checking and print out.
11167 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11168 Nehemiah processors. These extensions support AES encryption in hardware
11169 as well as RNG (though RNG support is currently disabled).
11171 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11173 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11177 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11179 *Andy Polyakov and a number of other people*
11181 * Improved PowerPC platform support. Most notably BIGNUM assembler
11182 implementation contributed by IBM.
11184 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11186 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11187 exponent rather than 'unsigned long'. There is a corresponding change to
11188 the new 'rsa_keygen' element of the RSA_METHOD structure.
11190 *Jelte Jansen, Geoff Thorpe*
11192 * Functionality for creating the initial serial number file is now
11193 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11195 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11196 number file to 1, which is bound to cause problems. To avoid
11197 the problems while respecting compatibility between different 0.9.7
11198 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11199 CA.pl for serial number initialization. With the new release 0.9.8,
11200 we can fix the problem directly in the 'ca' utility.)
11204 * Reduced header interdependencies by declaring more opaque objects in
11205 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11206 give fewer recursive includes, which could break lazy source code - so
11207 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11208 developers should define this symbol when building and using openssl to
11209 ensure they track the recommended behaviour, interfaces, [etc], but
11210 backwards-compatible behaviour prevails when this isn't defined.
11214 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11218 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11219 This will generate a random key of the appropriate length based on the
11220 cipher context. The EVP_CIPHER can provide its own random key generation
11221 routine to support keys of a specific form. This is used in the des and
11222 3des routines to generate a key of the correct parity. Update S/MIME
11223 code to use new functions and hence generate correct parity DES keys.
11224 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11225 valid (weak or incorrect parity).
11229 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11230 as looking them up. This is useful when the verified structure may contain
11231 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11232 present unless the new PKCS7_NO_CRL flag is asserted.
11236 * Extend ASN1 oid configuration module. It now additionally accepts the
11239 shortName = some long name, 1.2.3.4
11243 * Reimplemented the BN_CTX implementation. There is now no more static
11244 limitation on the number of variables it can handle nor the depth of the
11245 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11246 information can now expand as required, and rather than having a single
11247 static array of bignums, BN_CTX now uses a linked-list of such arrays
11248 allowing it to expand on demand whilst maintaining the usefulness of
11249 BN_CTX's "bundling".
11253 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11254 to allow all RSA operations to function using a single BN_CTX.
11258 * Preliminary support for certificate policy evaluation and checking. This
11259 is initially intended to pass the tests outlined in "Conformance Testing
11260 of Relying Party Client Certificate Path Processing Logic" v1.07.
11264 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11265 remained unused and not that useful. A variety of other little bignum
11266 tweaks and fixes have also been made continuing on from the audit (see
11271 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11272 associated ASN1, EVP and SSL functions and old ASN1 macros.
11276 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11277 and this should never fail. So the return value from the use of
11278 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11279 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11283 * BN_CTX_get() should return zero-valued bignums, providing the same
11284 initialised value as BN_new().
11286 *Geoff Thorpe, suggested by Ulf Möller*
11288 * Support for inhibitAnyPolicy certificate extension.
11292 * An audit of the BIGNUM code is underway, for which debugging code is
11293 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11294 is considered valid when processing BIGNUMs, and causes execution to
11295 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11296 further steps are taken to deliberately pollute unused data in BIGNUM
11297 structures to try and expose faulty code further on. For now, openssl will
11298 (in its default mode of operation) continue to tolerate the inconsistent
11299 forms that it has tolerated in the past, but authors and packagers should
11300 consider trying openssl and their own applications when compiled with
11301 these debugging symbols defined. It will help highlight potential bugs in
11302 their own code, and will improve the test coverage for OpenSSL itself. At
11303 some point, these tighter rules will become openssl's default to improve
11304 maintainability, though the assert()s and other overheads will remain only
11305 in debugging configurations. See bn.h for more details.
11307 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11309 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11310 that can only be obtained through BN_CTX_new() (which implicitly
11311 initialises it). The presence of this function only made it possible
11312 to overwrite an existing structure (and cause memory leaks).
11316 * Because of the callback-based approach for implementing LHASH as a
11317 template type, lh_insert() adds opaque objects to hash-tables and
11318 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11319 to clean up those corresponding objects before destroying the hash table
11320 (and losing the object pointers). So some over-zealous constifications in
11321 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11322 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11323 prototyped to have "const" restrictions on the object pointers they are
11324 given (and so aren't required to cast them away any more).
11328 * The tmdiff.h API was so ugly and minimal that our own timing utility
11329 (speed) prefers to use its own implementation. The two implementations
11330 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11331 its object type properly exposed (MS_TM) instead of casting to/from
11332 `char *`. This may still change yet if someone realises MS_TM and
11334 aren't necessarily the greatest nomenclatures - but this is what was used
11335 internally to the implementation so I've used that for now.
11339 * Ensure that deprecated functions do not get compiled when
11340 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11341 the self-tests were still using deprecated key-generation functions so
11342 these have been updated also.
11346 * Reorganise PKCS#7 code to separate the digest location functionality
11347 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11348 New function PKCS7_set_digest() to set the digest type for PKCS#7
11349 digestedData type. Add additional code to correctly generate the
11350 digestedData type and add support for this type in PKCS7 initialization
11355 * New function PKCS7_set0_type_other() this initializes a PKCS7
11356 structure of type "other".
11360 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11361 sure the loop does correctly stop and breaking ("division by zero")
11362 modulus operations are not performed. The (pre-generated) prime
11363 table crypto/bn/bn_prime.h was already correct, but it could not be
11364 re-generated on some platforms because of the "division by zero"
11365 situation in the script.
11367 *Ralf S. Engelschall*
11369 * Update support for ECC-based TLS ciphersuites according to
11370 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11371 SHA-1 now is only used for "small" curves (where the
11372 representation of a field element takes up to 24 bytes); for
11373 larger curves, the field element resulting from ECDH is directly
11374 used as premaster secret.
11376 *Douglas Stebila (Sun Microsystems Laboratories)*
11378 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11379 curve secp160r1 to the tests.
11381 *Douglas Stebila (Sun Microsystems Laboratories)*
11383 * Add the possibility to load symbols globally with DSO.
11385 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11387 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11388 control of the error stack.
11392 * Add support for STORE in ENGINE.
11396 * Add the STORE type. The intention is to provide a common interface
11397 to certificate and key stores, be they simple file-based stores, or
11398 HSM-type store, or LDAP stores, or...
11399 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11403 * Add a generic structure called OPENSSL_ITEM. This can be used to
11404 pass a list of arguments to any function as well as provide a way
11405 for a function to pass data back to the caller.
11409 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11410 works like BUF_strdup() but can be used to duplicate a portion of
11411 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11416 * Add the function sk_find_ex() which works like sk_find(), but will
11417 return an index to an element even if an exact match couldn't be
11418 found. The index is guaranteed to point at the element where the
11419 searched-for key would be inserted to preserve sorting order.
11423 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11424 takes an extra flags argument for optional functionality. Currently,
11425 the following flags are defined:
11427 OBJ_BSEARCH_VALUE_ON_NOMATCH
11428 This one gets OBJ_bsearch_ex() to return a pointer to the first
11429 element where the comparing function returns a negative or zero
11432 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11433 This one gets OBJ_bsearch_ex() to return a pointer to the first
11434 element where the comparing function returns zero. This is useful
11435 if there are more than one element where the comparing function
11440 * Make it possible to create self-signed certificates with 'openssl ca'
11441 in such a way that the self-signed certificate becomes part of the
11442 CA database and uses the same mechanisms for serial number generation
11443 as all other certificate signing. The new flag '-selfsign' enables
11444 this functionality. Adapt CA.sh and CA.pl.in.
11448 * Add functionality to check the public key of a certificate request
11449 against a given private. This is useful to check that a certificate
11450 request can be signed by that key (self-signing).
11454 * Make it possible to have multiple active certificates with the same
11455 subject in the CA index file. This is done only if the keyword
11456 'unique_subject' is set to 'no' in the main CA section (default
11457 if 'CA_default') of the configuration file. The value is saved
11458 with the database itself in a separate index attribute file,
11459 named like the index file with '.attr' appended to the name.
11463 * Generate multi-valued AVAs using '+' notation in config files for
11468 * Support for nameConstraints certificate extension.
11472 * Support for policyConstraints certificate extension.
11476 * Support for policyMappings certificate extension.
11480 * Make sure the default DSA_METHOD implementation only uses its
11481 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11482 and change its own handlers to be NULL so as to remove unnecessary
11483 indirection. This lets alternative implementations fallback to the
11484 default implementation more easily.
11488 * Support for directoryName in GeneralName related extensions
11493 * Make it possible to link applications using Makefile.shared.
11494 Make that possible even when linking against static libraries!
11498 * Support for single pass processing for S/MIME signing. This now
11499 means that S/MIME signing can be done from a pipe, in addition
11500 cleartext signing (multipart/signed type) is effectively streaming
11501 and the signed data does not need to be all held in memory.
11503 This is done with a new flag PKCS7_STREAM. When this flag is set
11504 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11505 is done after the data is output (and digests calculated) in
11506 SMIME_write_PKCS7().
11510 * Add full support for -rpath/-R, both in shared libraries and
11511 applications, at least on the platforms where it's known how
11516 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11517 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11518 will now compute a table of multiples of the generator that
11519 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11520 faster (notably in the case of a single point multiplication,
11521 scalar * generator).
11523 *Nils Larsch, Bodo Moeller*
11525 * IPv6 support for certificate extensions. The various extensions
11526 which use the IP:a.b.c.d can now take IPv6 addresses using the
11527 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11532 * Added an ENGINE that implements RSA by performing private key
11533 exponentiations with the GMP library. The conversions to and from
11534 GMP's mpz_t format aren't optimised nor are any montgomery forms
11535 cached, and on x86 it appears OpenSSL's own performance has caught up.
11536 However there are likely to be other architectures where GMP could
11537 provide a boost. This ENGINE is not built in by default, but it can be
11538 specified at Configure time and should be accompanied by the necessary
11539 linker additions, eg;
11540 ./config -DOPENSSL_USE_GMP -lgmp
11544 * "openssl engine" will not display ENGINE/DSO load failure errors when
11545 testing availability of engines with "-t" - the old behaviour is
11546 produced by increasing the feature's verbosity with "-tt".
11550 * ECDSA routines: under certain error conditions uninitialized BN objects
11551 could be freed. Solution: make sure initialization is performed early
11552 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11557 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11558 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11559 software implementations. For DSA and DH, parameter generation can
11560 also be overridden by providing the appropriate method callbacks.
11564 * Change the "progress" mechanism used in key-generation and
11565 primality testing to functions that take a new BN_GENCB pointer in
11566 place of callback/argument pairs. The new API functions have `_ex`
11567 postfixes and the older functions are reimplemented as wrappers for
11568 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11569 declarations of the old functions to help (graceful) attempts to
11570 migrate to the new functions. Also, the new key-generation API
11571 functions operate on a caller-supplied key-structure and return
11572 success/failure rather than returning a key or NULL - this is to
11573 help make "keygen" another member function of RSA_METHOD etc.
11575 Example for using the new callback interface:
11577 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11578 void *my_arg = ...;
11581 BN_GENCB_set(&my_cb, my_callback, my_arg);
11583 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11584 /* For the meaning of a, b in calls to my_callback(), see the
11585 * documentation of the function that calls the callback.
11586 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11587 * my_callback should return 1 if it wants BN_is_prime_ex()
11588 * to continue, or 0 to stop.
11593 * Change the ZLIB compression method to be stateful, and make it
11594 available to TLS with the number defined in
11595 draft-ietf-tls-compression-04.txt.
11599 * Add the ASN.1 structures and functions for CertificatePair, which
11600 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11602 CertificatePair ::= SEQUENCE {
11603 forward [0] Certificate OPTIONAL,
11604 reverse [1] Certificate OPTIONAL,
11605 -- at least one of the pair shall be present -- }
11607 Also implement the PEM functions to read and write certificate
11608 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11610 This needed to be defined, mostly for the sake of the LDAP
11611 attribute crossCertificatePair, but may prove useful elsewhere as
11616 * Make it possible to inhibit symlinking of shared libraries in
11617 Makefile.shared, for Cygwin's sake.
11621 * Extend the BIGNUM API by creating a function
11622 void BN_set_negative(BIGNUM *a, int neg);
11623 and a macro that behave like
11624 int BN_is_negative(const BIGNUM *a);
11626 to avoid the need to access 'a->neg' directly in applications.
11630 * Implement fast modular reduction for pseudo-Mersenne primes
11631 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11632 EC_GROUP_new_curve_GFp() will now automatically use this
11635 *Nils Larsch <nla@trustcenter.de>*
11637 * Add new lock type (CRYPTO_LOCK_BN).
11641 * Change the ENGINE framework to automatically load engines
11642 dynamically from specific directories unless they could be
11643 found to already be built in or loaded. Move all the
11644 current engines except for the cryptodev one to a new
11645 directory engines/.
11646 The engines in engines/ are built as shared libraries if
11647 the "shared" options was given to ./Configure or ./config.
11648 Otherwise, they are inserted in libcrypto.a.
11649 /usr/local/ssl/engines is the default directory for dynamic
11650 engines, but that can be overridden at configure time through
11651 the usual use of --prefix and/or --openssldir, and at run
11652 time with the environment variable OPENSSL_ENGINES.
11654 *Geoff Thorpe and Richard Levitte*
11656 * Add Makefile.shared, a helper makefile to build shared
11657 libraries. Adapt Makefile.org.
11661 * Add version info to Win32 DLLs.
11663 *Peter 'Luna' Runestig" <peter@runestig.com>*
11665 * Add new 'medium level' PKCS#12 API. Certificates and keys
11666 can be added using this API to created arbitrary PKCS#12
11667 files while avoiding the low-level API.
11669 New options to PKCS12_create(), key or cert can be NULL and
11670 will then be omitted from the output file. The encryption
11671 algorithm NIDs can be set to -1 for no encryption, the mac
11672 iteration count can be set to 0 to omit the mac.
11674 Enhance pkcs12 utility by making the -nokeys and -nocerts
11675 options work when creating a PKCS#12 file. New option -nomac
11676 to omit the mac, NONE can be set for an encryption algorithm.
11677 New code is modified to use the enhanced PKCS12_create()
11678 instead of the low-level API.
11682 * Extend ASN1 encoder to support indefinite length constructed
11683 encoding. This can output sequences tags and octet strings in
11684 this form. Modify pk7_asn1.c to support indefinite length
11685 encoding. This is experimental and needs additional code to
11686 be useful, such as an ASN1 bio and some enhanced streaming
11689 Extend template encode functionality so that tagging is passed
11690 down to the template encoder.
11694 * Let 'openssl req' fail if an argument to '-newkey' is not
11695 recognized instead of using RSA as a default.
11699 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11700 As these are not official, they are not included in "ALL";
11701 the "ECCdraft" ciphersuite group alias can be used to select them.
11703 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11705 * Add ECDH engine support.
11707 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11709 * Add ECDH in new directory crypto/ecdh/.
11711 *Douglas Stebila (Sun Microsystems Laboratories)*
11713 * Let BN_rand_range() abort with an error after 100 iterations
11714 without success (which indicates a broken PRNG).
11718 * Change BN_mod_sqrt() so that it verifies that the input value
11719 is really the square of the return value. (Previously,
11720 BN_mod_sqrt would show GIGO behaviour.)
11724 * Add named elliptic curves over binary fields from X9.62, SECG,
11725 and WAP/WTLS; add OIDs that were still missing.
11727 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11729 * Extend the EC library for elliptic curves over binary fields
11730 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11733 EC_GF2m_simple_method
11737 EC_GROUP_new_curve_GF2m
11738 EC_GROUP_set_curve_GF2m
11739 EC_GROUP_get_curve_GF2m
11740 EC_POINT_set_affine_coordinates_GF2m
11741 EC_POINT_get_affine_coordinates_GF2m
11742 EC_POINT_set_compressed_coordinates_GF2m
11744 Point compression for binary fields is disabled by default for
11745 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11748 As binary polynomials are represented as BIGNUMs, various members
11749 of the EC_GROUP and EC_POINT data structures can be shared
11750 between the implementations for prime fields and binary fields;
11751 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11752 are essentially identical to their `..._GFp` counterparts.
11753 (For simplicity, the `..._GFp` prefix has been dropped from
11754 various internal method names.)
11756 An internal 'field_div' method (similar to 'field_mul' and
11757 'field_sqr') has been added; this is used only for binary fields.
11759 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11761 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11762 through methods ('mul', 'precompute_mult').
11764 The generic implementations (now internally called 'ec_wNAF_mul'
11765 and 'ec_wNAF_precomputed_mult') remain the default if these
11766 methods are undefined.
11768 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11770 * New function EC_GROUP_get_degree, which is defined through
11771 EC_METHOD. For curves over prime fields, this returns the bit
11772 length of the modulus.
11774 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11776 * New functions EC_GROUP_dup, EC_POINT_dup.
11777 (These simply call ..._new and ..._copy).
11779 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11781 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11782 Polynomials are represented as BIGNUMs (where the sign bit is not
11783 used) in the following functions [macros]:
11786 BN_GF2m_sub [= BN_GF2m_add]
11787 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11788 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11789 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11791 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11792 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11793 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11794 BN_GF2m_cmp [= BN_ucmp]
11796 (Note that only the 'mod' functions are actually for fields GF(2^m).
11797 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11799 For some functions, an the irreducible polynomial defining a
11800 field can be given as an 'unsigned int[]' with strictly
11801 decreasing elements giving the indices of those bits that are set;
11802 i.e., p[] represents the polynomial
11803 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11805 p[0] > p[1] > ... > p[k] = 0.
11806 This applies to the following functions:
11809 BN_GF2m_mod_mul_arr
11810 BN_GF2m_mod_sqr_arr
11811 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11812 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11813 BN_GF2m_mod_exp_arr
11814 BN_GF2m_mod_sqrt_arr
11815 BN_GF2m_mod_solve_quad_arr
11819 Conversion can be performed by the following functions:
11824 bntest.c has additional tests for binary polynomial arithmetic.
11826 Two implementations for BN_GF2m_mod_div() are available.
11827 The default algorithm simply uses BN_GF2m_mod_inv() and
11828 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11829 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11830 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11832 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11834 * Add new error code 'ERR_R_DISABLED' that can be used when some
11835 functionality is disabled at compile-time.
11837 *Douglas Stebila <douglas.stebila@sun.com>*
11839 * Change default behaviour of 'openssl asn1parse' so that more
11840 information is visible when viewing, e.g., a certificate:
11842 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11843 mode the content of non-printable OCTET STRINGs is output in a
11844 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11845 avoid the appearance of a printable string.
11847 *Nils Larsch <nla@trustcenter.de>*
11849 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11851 EC_GROUP_set_asn1_flag()
11852 EC_GROUP_get_asn1_flag()
11853 EC_GROUP_set_point_conversion_form()
11854 EC_GROUP_get_point_conversion_form()
11855 These control ASN1 encoding details:
11856 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11857 has been set to OPENSSL_EC_NAMED_CURVE.
11858 - Points are encoded in uncompressed form by default; options for
11859 asn1_for are as for point2oct, namely
11860 POINT_CONVERSION_COMPRESSED
11861 POINT_CONVERSION_UNCOMPRESSED
11862 POINT_CONVERSION_HYBRID
11864 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11866 EC_GROUP_set_seed()
11867 EC_GROUP_get0_seed()
11868 EC_GROUP_get_seed_len()
11869 This is used only for ASN1 purposes (so far).
11871 *Nils Larsch <nla@trustcenter.de>*
11873 * Add 'field_type' member to EC_METHOD, which holds the NID
11874 of the appropriate field type OID. The new function
11875 EC_METHOD_get_field_type() returns this value.
11877 *Nils Larsch <nla@trustcenter.de>*
11880 EC_POINT_point2bn()
11881 EC_POINT_bn2point()
11882 EC_POINT_point2hex()
11883 EC_POINT_hex2point()
11884 providing useful interfaces to EC_POINT_point2oct() and
11885 EC_POINT_oct2point().
11887 *Nils Larsch <nla@trustcenter.de>*
11889 * Change internals of the EC library so that the functions
11890 EC_GROUP_set_generator()
11891 EC_GROUP_get_generator()
11892 EC_GROUP_get_order()
11893 EC_GROUP_get_cofactor()
11894 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11895 to methods, which would lead to unnecessary code duplication when
11896 adding different types of curves.
11898 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11900 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11901 arithmetic, and such that modified wNAFs are generated
11902 (which avoid length expansion in many cases).
11906 * Add a function EC_GROUP_check_discriminant() (defined via
11907 EC_METHOD) that verifies that the curve discriminant is non-zero.
11909 Add a function EC_GROUP_check() that makes some sanity tests
11910 on a EC_GROUP, its generator and order. This includes
11911 EC_GROUP_check_discriminant().
11913 *Nils Larsch <nla@trustcenter.de>*
11915 * Add ECDSA in new directory crypto/ecdsa/.
11917 Add applications 'openssl ecparam' and 'openssl ecdsa'
11918 (these are based on 'openssl dsaparam' and 'openssl dsa').
11920 ECDSA support is also included in various other files across the
11921 library. Most notably,
11922 - 'openssl req' now has a '-newkey ecdsa:file' option;
11923 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11924 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11925 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11926 them suitable for ECDSA where domain parameters must be
11927 extracted before the specific public key;
11928 - ECDSA engine support has been added.
11930 *Nils Larsch <nla@trustcenter.de>*
11932 * Include some named elliptic curves, and add OIDs from X9.62,
11933 SECG, and WAP/WTLS. Each curve can be obtained from the new
11935 EC_GROUP_new_by_curve_name(),
11936 and the list of available named curves can be obtained with
11937 EC_get_builtin_curves().
11938 Also add a 'curve_name' member to EC_GROUP objects, which can be
11940 EC_GROUP_set_curve_name()
11941 EC_GROUP_get_curve_name()
11943 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11945 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11946 was actually never needed) and in BN_mul(). The removal in BN_mul()
11947 required a small change in bn_mul_part_recursive() and the addition
11948 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11949 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11950 bn_sub_words() and bn_add_words() except they take arrays with
11955 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11957 * Cleanse PEM buffers before freeing them since they may contain
11960 *Benjamin Bennett <ben@psc.edu>*
11962 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11963 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11964 authentication-only ciphersuites.
11968 * Since AES128 and AES256 share a single mask bit in the logic of
11969 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11970 kludge to work properly if AES128 is available and AES256 isn't.
11974 * Expand security boundary to match 1.1.1 module.
11978 * Remove redundant features: hash file source, editing of test vectors
11979 modify fipsld to use external fips_premain.c signature.
11983 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11984 run algorithm test programs.
11988 * Make algorithm test programs more tolerant of whitespace.
11992 * Have SSL/TLS server implementation tolerate "mismatched" record
11993 protocol version while receiving ClientHello even if the
11994 ClientHello is fragmented. (The server can't insist on the
11995 particular protocol version it has chosen before the ServerHello
11996 message has informed the client about his choice.)
12000 * Load error codes if they are not already present instead of using a
12001 static variable. This allows them to be cleanly unloaded and reloaded.
12005 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12007 * Introduce limits to prevent malicious keys being able to
12008 cause a denial of service. ([CVE-2006-2940])
12010 *Steve Henson, Bodo Moeller*
12012 * Fix ASN.1 parsing of certain invalid structures that can result
12013 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12015 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12016 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12018 * Fix SSL client code which could crash if connecting to a
12019 malicious SSLv2 server. ([CVE-2006-4343])
12021 *Tavis Ormandy and Will Drewry, Google Security Team*
12023 * Change ciphersuite string processing so that an explicit
12024 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12025 will no longer include "AES128-SHA"), and any other similar
12026 ciphersuite (same bitmap) from *other* protocol versions (so that
12027 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12028 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12029 changes from 0.9.8b and 0.9.8d.
12033 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12035 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12036 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12038 * Change the Unix randomness entropy gathering to use poll() when
12039 possible instead of select(), since the latter has some
12040 undesirable limitations.
12042 *Darryl Miles via Richard Levitte and Bodo Moeller*
12044 * Disable rogue ciphersuites:
12046 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12047 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12048 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12050 The latter two were purportedly from
12051 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12054 Also deactivate the remaining ciphersuites from
12055 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12056 unofficial, and the ID has long expired.
12060 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12061 dual-core machines) and other potential thread-safety issues.
12065 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12067 * Adapt fipsld and the build system to link against the validated FIPS
12068 module in FIPS mode.
12072 * Fixes for VC++ 2005 build under Windows.
12076 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12077 from a Windows bash shell such as MSYS. It is autodetected from the
12078 "config" script when run from a VC++ environment. Modify standard VC++
12079 build to use fipscanister.o from the GNU make build.
12083 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12085 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12086 The value now differs depending on if you build for FIPS or not.
12087 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12088 safely run with a non-FIPSed libcrypto, as it may crash because of
12089 the difference induced by this change.
12093 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12095 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12096 (part of SSL_OP_ALL). This option used to disable the
12097 countermeasure against man-in-the-middle protocol-version
12098 rollback in the SSL 2.0 server implementation, which is a bad
12099 idea. ([CVE-2005-2969])
12101 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12102 for Information Security, National Institute of Advanced Industrial
12103 Science and Technology [AIST, Japan)]*
12105 * Minimal support for X9.31 signatures and PSS padding modes. This is
12106 mainly for FIPS compliance and not fully integrated at this stage.
12110 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12111 the exponentiation using a fixed-length exponent. (Otherwise,
12112 the information leaked through timing could expose the secret key
12113 after many signatures; cf. Bleichenbacher's attack on DSA with
12118 * Make a new fixed-window mod_exp implementation the default for
12119 RSA, DSA, and DH private-key operations so that the sequence of
12120 squares and multiplies and the memory access pattern are
12121 independent of the particular secret key. This will mitigate
12122 cache-timing and potential related attacks.
12124 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12125 and this is automatically used by BN_mod_exp_mont() if the new flag
12126 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12127 will use this BN flag for private exponents unless the flag
12128 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12129 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12131 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12133 * Change the client implementation for SSLv23_method() and
12134 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12135 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12136 (Previously, the SSL 2.0 backwards compatible Client Hello
12137 message format would be used even with SSL_OP_NO_SSLv2.)
12141 * Add support for smime-type MIME parameter in S/MIME messages which some
12146 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12147 a threadsafe manner. Modify rsa code to use new function and add calls
12148 to dsa and dh code (which had race conditions before).
12152 * Include the fixed error library code in the C error file definitions
12153 instead of fixing them up at runtime. This keeps the error code
12154 structures constant.
12158 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12160 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12163 * Fixes for newer kerberos headers. NB: the casts are needed because
12164 the 'length' field is signed on one version and unsigned on another
12165 with no (?) obvious way to tell the difference, without these VC++
12166 complains. Also the "definition" of FAR (blank) is no longer included
12167 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12168 some needed definitions.
12172 * Undo Cygwin change.
12176 * Added support for proxy certificates according to RFC 3820.
12177 Because they may be a security thread to unaware applications,
12178 they must be explicitly allowed in run-time. See
12179 docs/HOWTO/proxy_certificates.txt for further information.
12183 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12185 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12186 server and client random values. Previously
12187 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12188 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12190 This change has negligible security impact because:
12192 1. Server and client random values still have 24 bytes of pseudo random
12195 2. Server and client random values are sent in the clear in the initial
12198 3. The master secret is derived using the premaster secret (48 bytes in
12199 size for static RSA ciphersuites) as well as client server and random
12202 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12205 *Stephen Henson, reported by UK NISCC*
12207 * Use Windows randomness collection on Cygwin.
12211 * Fix hang in EGD/PRNGD query when communication socket is closed
12212 prematurely by EGD/PRNGD.
12214 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12216 * Prompt for pass phrases when appropriate for PKCS12 input format.
12220 * Back-port of selected performance improvements from development
12221 branch, as well as improved support for PowerPC platforms.
12225 * Add lots of checks for memory allocation failure, error codes to indicate
12226 failure and freeing up memory if a failure occurs.
12228 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12230 * Add new -passin argument to dgst.
12234 * Perform some character comparisons of different types in X509_NAME_cmp:
12235 this is needed for some certificates that re-encode DNs into UTF8Strings
12236 (in violation of RFC3280) and can't or won't issue name rollover
12241 * Make an explicit check during certificate validation to see that
12242 the CA setting in each certificate on the chain is correct. As a
12243 side effect always do the following basic checks on extensions,
12244 not just when there's an associated purpose to the check:
12246 - if there is an unhandled critical extension (unless the user
12247 has chosen to ignore this fault)
12248 - if the path length has been exceeded (if one is set at all)
12249 - that certain extensions fit the associated purpose (if one has
12254 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12256 * Avoid a race condition when CRLs are checked in a multi threaded
12257 environment. This would happen due to the reordering of the revoked
12258 entries during signature checking and serial number lookup. Now the
12259 encoding is cached and the serial number sort performed under a lock.
12260 Add new STACK function sk_is_sorted().
12264 * Add Delta CRL to the extension code.
12268 * Various fixes to s3_pkt.c so alerts are sent properly.
12270 *David Holmes <d.holmes@f5.com>*
12272 * Reduce the chances of duplicate issuer name and serial numbers (in
12273 violation of RFC3280) using the OpenSSL certificate creation utilities.
12274 This is done by creating a random 64 bit value for the initial serial
12275 number when a serial number file is created or when a self signed
12276 certificate is created using 'openssl req -x509'. The initial serial
12277 number file is created using 'openssl x509 -next_serial' in CA.pl
12278 rather than being initialized to 1.
12282 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12284 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12285 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12287 *Joe Orton, Steve Henson*
12289 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12292 *Joe Orton, Steve Henson*
12294 * Make it possible to have multiple active certificates with the same
12295 subject in the CA index file. This is done only if the keyword
12296 'unique_subject' is set to 'no' in the main CA section (default
12297 if 'CA_default') of the configuration file. The value is saved
12298 with the database itself in a separate index attribute file,
12299 named like the index file with '.attr' appended to the name.
12303 * X509 verify fixes. Disable broken certificate workarounds when
12304 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12305 keyUsage extension present. Don't accept CRLs with unhandled critical
12306 extensions: since verify currently doesn't process CRL extensions this
12307 rejects a CRL with *any* critical extensions. Add new verify error codes
12312 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12313 A clarification of RFC2560 will require the use of OCTET STRINGs and
12314 some implementations cannot handle the current raw format. Since OpenSSL
12315 copies and compares OCSP nonces as opaque blobs without any attempt at
12316 parsing them this should not create any compatibility issues.
12320 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12321 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12322 this HMAC (and other) operations are several times slower than OpenSSL
12327 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12329 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12331 * Use the correct content when signing type "other".
12335 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12337 * Fix various bugs revealed by running the NISCC test suite:
12339 Stop out of bounds reads in the ASN1 code when presented with
12340 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12342 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12344 If verify callback ignores invalid public key errors don't try to check
12345 certificate signature with the NULL public key.
12349 * New -ignore_err option in ocsp application to stop the server
12350 exiting on the first error in a request.
12354 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12355 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12360 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12361 extra data after the compression methods not only for TLS 1.0
12362 but also for SSL 3.0 (as required by the specification).
12364 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12366 * Change X509_certificate_type() to mark the key as exported/exportable
12367 when it's 512 *bits* long, not 512 bytes.
12371 * Change AES_cbc_encrypt() so it outputs exact multiple of
12372 blocks during encryption.
12376 * Various fixes to base64 BIO and non blocking I/O. On write
12377 flushes were not handled properly if the BIO retried. On read
12378 data was not being buffered properly and had various logic bugs.
12379 This also affects blocking I/O when the data being decoded is a
12384 * Various S/MIME bugfixes and compatibility changes:
12385 output correct application/pkcs7 MIME type if
12386 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12387 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12388 of files as .eml work). Correctly handle very long lines in MIME
12393 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12395 * Countermeasure against the Klima-Pokorny-Rosa extension of
12396 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12397 a protocol version number mismatch like a decryption error
12398 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12402 * Turn on RSA blinding by default in the default implementation
12403 to avoid a timing attack. Applications that don't want it can call
12404 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12405 They would be ill-advised to do so in most cases.
12407 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12409 * Change RSA blinding code so that it works when the PRNG is not
12410 seeded (in this case, the secret RSA exponent is abused as
12411 an unpredictable seed -- if it is not unpredictable, there
12412 is no point in blinding anyway). Make RSA blinding thread-safe
12413 by remembering the creator's thread ID in rsa->blinding and
12414 having all other threads use local one-time blinding factors
12415 (this requires more computation than sharing rsa->blinding, but
12416 avoids excessive locking; and if an RSA object is not shared
12417 between threads, blinding will still be very fast).
12421 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12422 ENGINE as defaults for all supported algorithms irrespective of
12423 the 'flags' parameter. 'flags' is now honoured, so applications
12424 should make sure they are passing it correctly.
12428 * Target "mingw" now allows native Windows code to be generated in
12429 the Cygwin environment as well as with the MinGW compiler.
12433 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12435 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12436 via timing by performing a MAC computation even if incorrect
12437 block cipher padding has been found. This is a countermeasure
12438 against active attacks where the attacker has to distinguish
12439 between bad padding and a MAC verification error. ([CVE-2003-0078])
12441 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12442 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12443 Martin Vuagnoux (EPFL, Ilion)*
12445 * Make the no-err option work as intended. The intention with no-err
12446 is not to have the whole error stack handling routines removed from
12447 libcrypto, it's only intended to remove all the function name and
12448 reason texts, thereby removing some of the footprint that may not
12449 be interesting if those errors aren't displayed anyway.
12451 NOTE: it's still possible for any application or module to have its
12452 own set of error texts inserted. The routines are there, just not
12453 used by default when no-err is given.
12457 * Add support for FreeBSD on IA64.
12459 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12461 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12462 Kerberos function mit_des_cbc_cksum(). Before this change,
12463 the value returned by DES_cbc_cksum() was like the one from
12464 mit_des_cbc_cksum(), except the bytes were swapped.
12466 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12468 * Allow an application to disable the automatic SSL chain building.
12469 Before this a rather primitive chain build was always performed in
12470 ssl3_output_cert_chain(): an application had no way to send the
12471 correct chain if the automatic operation produced an incorrect result.
12473 Now the chain builder is disabled if either:
12475 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12477 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12479 The reasoning behind this is that an application would not want the
12480 auto chain building to take place if extra chain certificates are
12481 present and it might also want a means of sending no additional
12482 certificates (for example the chain has two certificates and the
12487 * Add the possibility to build without the ENGINE framework.
12489 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12491 * Under Win32 gmtime() can return NULL: check return value in
12492 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12496 * DSA routines: under certain error conditions uninitialized BN objects
12497 could be freed. Solution: make sure initialization is performed early
12498 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12499 Nils Larsch <nla@trustcenter.de> via PR#459)
12503 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12504 checked on reconnect on the client side, therefore session resumption
12505 could still fail with a "ssl session id is different" error. This
12506 behaviour is masked when SSL_OP_ALL is used due to
12507 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12508 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12509 followup to PR #377.
12513 * IA-32 assembler support enhancements: unified ELF targets, support
12514 for SCO/Caldera platforms, fix for Cygwin shared build.
12518 * Add support for FreeBSD on sparc64. As a consequence, support for
12519 FreeBSD on non-x86 processors is separate from x86 processors on
12520 the config script, much like the NetBSD support.
12522 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12524 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12526 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12529 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12530 code (06) was taken as the first octet of the session ID and the last
12531 octet was ignored consequently. As a result SSLv2 client side session
12532 caching could not have worked due to the session ID mismatch between
12534 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12539 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12540 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12545 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12546 seems that in spite of existing for more than a year, many application
12547 author have done nothing to provide the necessary callbacks, which
12548 means that this particular engine will not work properly anywhere.
12549 This is a very unfortunate situation which forces us, in the name
12550 of usability, to give the hw_ncipher.c a static lock, which is part
12552 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12553 appear in 0.9.8 or later. We EXPECT application authors to have
12554 dealt properly with this when 0.9.8 is released (unless we actually
12555 make such changes in the libcrypto locking code that changes will
12556 have to be made anyway).
12560 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12561 octets have been read, EOF or an error occurs. Without this change
12562 some truncated ASN1 structures will not produce an error.
12566 * Disable Heimdal support, since it hasn't been fully implemented.
12567 Still give the possibility to force the use of Heimdal, but with
12568 warnings and a request that patches get sent to openssl-dev.
12572 * Add the VC-CE target, introduce the WINCE sysname, and add
12573 INSTALL.WCE and appropriate conditionals to make it build.
12575 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12577 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12578 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12579 edit numbers of the version.
12581 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12583 * Introduce safe string copy and catenation functions
12584 (BUF_strlcpy() and BUF_strlcat()).
12586 *Ben Laurie (CHATS) and Richard Levitte*
12588 * Avoid using fixed-size buffers for one-line DNs.
12590 *Ben Laurie (CHATS)*
12592 * Add BUF_MEM_grow_clean() to avoid information leakage when
12593 resizing buffers containing secrets, and use where appropriate.
12595 *Ben Laurie (CHATS)*
12597 * Avoid using fixed size buffers for configuration file location.
12599 *Ben Laurie (CHATS)*
12601 * Avoid filename truncation for various CA files.
12603 *Ben Laurie (CHATS)*
12605 * Use sizeof in preference to magic numbers.
12607 *Ben Laurie (CHATS)*
12609 * Avoid filename truncation in cert requests.
12611 *Ben Laurie (CHATS)*
12613 * Add assertions to check for (supposedly impossible) buffer
12616 *Ben Laurie (CHATS)*
12618 * Don't cache truncated DNS entries in the local cache (this could
12619 potentially lead to a spoofing attack).
12621 *Ben Laurie (CHATS)*
12623 * Fix various buffers to be large enough for hex/decimal
12624 representations in a platform independent manner.
12626 *Ben Laurie (CHATS)*
12628 * Add CRYPTO_realloc_clean() to avoid information leakage when
12629 resizing buffers containing secrets, and use where appropriate.
12631 *Ben Laurie (CHATS)*
12633 * Add BIO_indent() to avoid much slightly worrying code to do
12636 *Ben Laurie (CHATS)*
12638 * Convert sprintf()/BIO_puts() to BIO_printf().
12640 *Ben Laurie (CHATS)*
12642 * buffer_gets() could terminate with the buffer only half
12645 *Ben Laurie (CHATS)*
12647 * Add assertions to prevent user-supplied crypto functions from
12648 overflowing internal buffers by having large block sizes, etc.
12650 *Ben Laurie (CHATS)*
12652 * New OPENSSL_assert() macro (similar to assert(), but enabled
12655 *Ben Laurie (CHATS)*
12657 * Eliminate unused copy of key in RC4.
12659 *Ben Laurie (CHATS)*
12661 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12663 *Ben Laurie (CHATS)*
12665 * Fix off-by-one error in EGD path.
12667 *Ben Laurie (CHATS)*
12669 * If RANDFILE path is too long, ignore instead of truncating.
12671 *Ben Laurie (CHATS)*
12673 * Eliminate unused and incorrectly sized X.509 structure
12676 *Ben Laurie (CHATS)*
12678 * Eliminate unused and dangerous function knumber().
12680 *Ben Laurie (CHATS)*
12682 * Eliminate unused and dangerous structure, KSSL_ERR.
12684 *Ben Laurie (CHATS)*
12686 * Protect against overlong session ID context length in an encoded
12687 session object. Since these are local, this does not appear to be
12690 *Ben Laurie (CHATS)*
12692 * Change from security patch (see 0.9.6e below) that did not affect
12693 the 0.9.6 release series:
12695 Remote buffer overflow in SSL3 protocol - an attacker could
12696 supply an oversized master key in Kerberos-enabled versions.
12699 *Ben Laurie (CHATS)*
12701 * Change the SSL kerb5 codes to match RFC 2712.
12705 * Make -nameopt work fully for req and add -reqopt switch.
12707 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12709 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12711 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12713 * Make sure tests can be performed even if the corresponding algorithms
12714 have been removed entirely. This was also the last step to make
12715 OpenSSL compilable with DJGPP under all reasonable conditions.
12717 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12719 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12720 to allow version independent disabling of normally unselected ciphers,
12721 which may be activated as a side-effect of selecting a single cipher.
12723 (E.g., cipher list string "RSA" enables ciphersuites that are left
12724 out of "ALL" because they do not provide symmetric encryption.
12725 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12727 *Lutz Jaenicke, Bodo Moeller*
12729 * Add appropriate support for separate platform-dependent build
12730 directories. The recommended way to make a platform-dependent
12731 build directory is the following (tested on Linux), maybe with
12734 # Place yourself outside of the OpenSSL source tree. In
12735 # this example, the environment variable OPENSSL_SOURCE
12736 # is assumed to contain the absolute OpenSSL source directory.
12737 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12738 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12739 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12740 mkdir -p `dirname $F`
12741 ln -s $OPENSSL_SOURCE/$F $F
12744 To be absolutely sure not to disturb the source tree, a "make clean"
12745 is a good thing. If it isn't successful, don't worry about it,
12746 it probably means the source directory is very clean.
12750 * Make sure any ENGINE control commands make local copies of string
12751 pointers passed to them whenever necessary. Otherwise it is possible
12752 the caller may have overwritten (or deallocated) the original string
12753 data when a later ENGINE operation tries to use the stored values.
12755 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12757 * Improve diagnostics in file reading and command-line digests.
12759 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12761 * Add AES modes CFB and OFB to the object database. Correct an
12762 error in AES-CFB decryption.
12766 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12767 allows existing EVP_CIPHER_CTX structures to be reused after
12768 calling `EVP_*Final()`. This behaviour is used by encryption
12769 BIOs and some applications. This has the side effect that
12770 applications must explicitly clean up cipher contexts with
12771 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12775 * Check the values of dna and dnb in bn_mul_recursive before calling
12776 bn_mul_comba (a non zero value means the a or b arrays do not contain
12777 n2 elements) and fallback to bn_mul_normal if either is not zero.
12781 * Fix escaping of non-ASCII characters when using the -subj option
12782 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12786 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12787 form for "surname", serialNumber has no short form.
12788 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12789 therefore remove "mail" short name for "internet 7".
12790 The OID for unique identifiers in X509 certificates is
12791 x500UniqueIdentifier, not uniqueIdentifier.
12792 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12796 * Add an "init" command to the ENGINE config module and auto initialize
12797 ENGINEs. Without any "init" command the ENGINE will be initialized
12798 after all ctrl commands have been executed on it. If init=1 the
12799 ENGINE is initialized at that point (ctrls before that point are run
12800 on the uninitialized ENGINE and after on the initialized one). If
12801 init=0 then the ENGINE will not be initialized at all.
12805 * Fix the 'app_verify_callback' interface so that the user-defined
12806 argument is actually passed to the callback: In the
12807 SSL_CTX_set_cert_verify_callback() prototype, the callback
12808 declaration has been changed from
12811 int (*cb)(X509_STORE_CTX *,void *);
12812 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12813 i=s->ctx->app_verify_callback(&ctx)
12814 has been changed into
12815 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12817 To update applications using SSL_CTX_set_cert_verify_callback(),
12818 a dummy argument can be added to their callback functions.
12820 *D. K. Smetters <smetters@parc.xerox.com>*
12822 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12824 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12826 * Add and OPENSSL_LOAD_CONF define which will cause
12827 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12828 This allows older applications to transparently support certain
12829 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12830 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12831 load the config file and OPENSSL_add_all_algorithms_conf() which will
12832 always load it have also been added.
12836 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12837 Adjust NIDs and EVP layer.
12839 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12841 * Config modules support in openssl utility.
12843 Most commands now load modules from the config file,
12844 though in a few (such as version) this isn't done
12845 because it couldn't be used for anything.
12847 In the case of ca and req the config file used is
12848 the same as the utility itself: that is the -config
12849 command line option can be used to specify an
12854 * Move default behaviour from OPENSSL_config(). If appname is NULL
12855 use "openssl_conf" if filename is NULL use default openssl config file.
12859 * Add an argument to OPENSSL_config() to allow the use of an alternative
12860 config section name. Add a new flag to tolerate a missing config file
12861 and move code to CONF_modules_load_file().
12865 * Support for crypto accelerator cards from Accelerated Encryption
12866 Processing, www.aep.ie. (Use engine 'aep')
12867 The support was copied from 0.9.6c [engine] and adapted/corrected
12868 to work with the new engine framework.
12870 *AEP Inc. and Richard Levitte*
12872 * Support for SureWare crypto accelerator cards from Baltimore
12873 Technologies. (Use engine 'sureware')
12874 The support was copied from 0.9.6c [engine] and adapted
12875 to work with the new engine framework.
12879 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12880 make the newer ENGINE framework commands for the CHIL engine work.
12882 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12884 * Make it possible to produce shared libraries on ReliantUNIX.
12886 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12888 * Add the configuration target debug-linux-ppro.
12889 Make 'openssl rsa' use the general key loading routines
12890 implemented in `apps.c`, and make those routines able to
12891 handle the key format FORMAT_NETSCAPE and the variant
12894 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12896 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12898 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12900 * Add -keyform to rsautl, and document -engine.
12902 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12904 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12905 BIO_R_NO_SUCH_FILE error code rather than the generic
12906 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12910 * Add new functions
12911 ERR_peek_last_error
12912 ERR_peek_last_error_line
12913 ERR_peek_last_error_line_data.
12914 These are similar to
12916 ERR_peek_error_line
12917 ERR_peek_error_line_data,
12918 but report on the latest error recorded rather than the first one
12919 still in the error queue.
12921 *Ben Laurie, Bodo Moeller*
12923 * default_algorithms option in ENGINE config module. This allows things
12925 default_algorithms = ALL
12926 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12930 * Preliminary ENGINE config module.
12934 * New experimental application configuration code.
12938 * Change the AES code to follow the same name structure as all other
12939 symmetric ciphers, and behave the same way. Move everything to
12940 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12942 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12944 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12946 *Ben Laurie and Theo de Raadt*
12948 * Add option to output public keys in req command.
12950 *Massimiliano Pala madwolf@openca.org*
12952 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12953 (up to about 10% better than before for P-192 and P-224).
12957 * New functions/macros
12959 SSL_CTX_set_msg_callback(ctx, cb)
12960 SSL_CTX_set_msg_callback_arg(ctx, arg)
12961 SSL_set_msg_callback(ssl, cb)
12962 SSL_set_msg_callback_arg(ssl, arg)
12964 to request calling a callback function
12966 void cb(int write_p, int version, int content_type,
12967 const void *buf, size_t len, SSL *ssl, void *arg)
12969 whenever a protocol message has been completely received
12970 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12971 protocol version according to which the SSL library interprets
12972 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12973 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12974 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12975 specification (change_cipher_spec(20), alert(21), handshake(22)).
12976 'buf' and 'len' point to the actual message, 'ssl' to the
12977 SSL object, and 'arg' is the application-defined value set by
12978 SSL[_CTX]_set_msg_callback_arg().
12980 'openssl s_client' and 'openssl s_server' have new '-msg' options
12981 to enable a callback that displays all protocol messages.
12985 * Change the shared library support so shared libraries are built as
12986 soon as the corresponding static library is finished, and thereby get
12987 openssl and the test programs linked against the shared library.
12988 This still only happens when the keyword "shard" has been given to
12989 the configuration scripts.
12991 NOTE: shared library support is still an experimental thing, and
12992 backward binary compatibility is still not guaranteed.
12994 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12996 * Add support for Subject Information Access extension.
12998 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13000 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13001 additional bytes when new memory had to be allocated, not just
13002 when reusing an existing buffer.
13006 * New command line and configuration option 'utf8' for the req command.
13007 This allows field values to be specified as UTF8 strings.
13011 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13012 runs for the former and machine-readable output for the latter.
13016 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13017 of the e-mail address in the DN (i.e., it will go into a certificate
13018 extension only). The new configuration file option 'email_in_dn = no'
13019 has the same effect.
13021 *Massimiliano Pala madwolf@openca.org*
13023 * Change all functions with names starting with `des_` to be starting
13024 with `DES_` instead. Add wrappers that are compatible with libdes,
13025 but are named `_ossl_old_des_*`. Finally, add macros that map the
13026 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13027 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13028 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13031 Since we provide two compatibility mappings, the user needs to
13032 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13033 compatibility is desired. The default (i.e., when that macro
13034 isn't defined) is OpenSSL 0.9.6c compatibility.
13036 There are also macros that enable and disable the support of old
13037 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13038 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13039 are defined, the default will apply: to support the old des routines.
13041 In either case, one must include openssl/des.h to get the correct
13042 definitions. Do not try to just include openssl/des_old.h, that
13045 NOTE: This is a major break of an old API into a new one. Software
13046 authors are encouraged to switch to the `DES_` style functions. Some
13047 time in the future, des_old.h and the libdes compatibility functions
13048 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13049 default), and then completely removed.
13053 * Test for certificates which contain unsupported critical extensions.
13054 If such a certificate is found during a verify operation it is
13055 rejected by default: this behaviour can be overridden by either
13056 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13057 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13058 X509_supported_extension() has also been added which returns 1 if a
13059 particular extension is supported.
13063 * Modify the behaviour of EVP cipher functions in similar way to digests
13064 to retain compatibility with existing code.
13068 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13069 compatibility with existing code. In particular the 'ctx' parameter does
13070 not have to be to be initialized before the call to EVP_DigestInit() and
13071 it is tidied up after a call to EVP_DigestFinal(). New function
13072 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13073 EVP_MD_CTX_copy() changed to not require the destination to be
13074 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13075 requires the destination to be valid.
13077 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13078 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13082 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13083 so that complete 'Handshake' protocol structures are kept in memory
13084 instead of overwriting 'msg_type' and 'length' with 'body' data.
13088 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13090 *Massimo Santin via Richard Levitte*
13092 * Major restructuring to the underlying ENGINE code. This includes
13093 reduction of linker bloat, separation of pure "ENGINE" manipulation
13094 (initialisation, etc) from functionality dealing with implementations
13095 of specific crypto interfaces. This change also introduces integrated
13096 support for symmetric ciphers and digest implementations - so ENGINEs
13097 can now accelerate these by providing EVP_CIPHER and EVP_MD
13098 implementations of their own. This is detailed in
13099 [crypto/engine/README.md](crypto/engine/README.md)
13100 as it couldn't be adequately described here. However, there are a few
13101 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13102 were changed in the original introduction of ENGINE code have now
13103 reverted back - the hooking from this code to ENGINE is now a good
13104 deal more passive and at run-time, operations deal directly with
13105 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13106 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13107 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13108 they were not being used by the framework as there is no concept of a
13109 BIGNUM_METHOD and they could not be generalised to the new
13110 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13111 ENGINE_cpy() has been removed as it cannot be consistently defined in
13116 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13120 * Change mkdef.pl to sort symbols that get the same entry number,
13121 and make sure the automatically generated functions `ERR_load_*`
13122 become part of libeay.num as well.
13126 * New function SSL_renegotiate_pending(). This returns true once
13127 renegotiation has been requested (either SSL_renegotiate() call
13128 or HelloRequest/ClientHello received from the peer) and becomes
13129 false once a handshake has been completed.
13130 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13131 sends a HelloRequest, but does not ensure that a handshake takes
13132 place. SSL_renegotiate_pending() is useful for checking if the
13133 client has followed the request.)
13137 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13138 By default, clients may request session resumption even during
13139 renegotiation (if session ID contexts permit); with this option,
13140 session resumption is possible only in the first handshake.
13142 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13143 more bits available for options that should not be part of
13144 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13148 * Add some demos for certificate and certificate request creation.
13152 * Make maximum certificate chain size accepted from the peer application
13153 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13154 "Douglas E. Engert" <deengert@anl.gov>.
13158 * Add support for shared libraries for Unixware-7
13159 (Boyd Lynn Gerber <gerberb@zenez.com>).
13163 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13164 be done prior to destruction. Use this to unload error strings from
13165 ENGINEs that load their own error strings. NB: This adds two new API
13166 functions to "get" and "set" this destroy handler in an ENGINE.
13170 * Alter all existing ENGINE implementations (except "openssl" and
13171 "openbsd") to dynamically instantiate their own error strings. This
13172 makes them more flexible to be built both as statically-linked ENGINEs
13173 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13174 Also, add stub code to each that makes building them as self-contained
13175 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13179 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13180 implementations into applications that are completely implemented in
13181 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13182 commands that can be used to configure what shared-library to load and
13183 to control aspects of the way it is handled. Also, made an update to
13184 the [README-Engine.md](README-Engine.md) file
13185 that brings its information up-to-date and
13186 provides some information and instructions on the "dynamic" ENGINE
13187 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13191 * Make it possible to unload ranges of ERR strings with a new
13192 "ERR_unload_strings" function.
13196 * Add a copy() function to EVP_MD.
13200 * Make EVP_MD routines take a context pointer instead of just the
13201 md_data void pointer.
13205 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13206 that the digest can only process a single chunk of data
13207 (typically because it is provided by a piece of
13208 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13209 is only going to provide a single chunk of data, and hence the
13210 framework needn't accumulate the data for oneshot drivers.
13214 * As with "ERR", make it possible to replace the underlying "ex_data"
13215 functions. This change also alters the storage and management of global
13216 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13217 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13218 index counters. The API functions that use this state have been changed
13219 to take a "class_index" rather than pointers to the class's local STACK
13220 and counter, and there is now an API function to dynamically create new
13221 classes. This centralisation allows us to (a) plug a lot of the
13222 thread-safety problems that existed, and (b) makes it possible to clean
13223 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13224 such data would previously have always leaked in application code and
13225 workarounds were in place to make the memory debugging turn a blind eye
13226 to it. Application code that doesn't use this new function will still
13227 leak as before, but their memory debugging output will announce it now
13228 rather than letting it slide.
13230 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13231 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13232 has a return value to indicate success or failure.
13236 * Make it possible to replace the underlying "ERR" functions such that the
13237 global state (2 LHASH tables and 2 locks) is only used by the "default"
13238 implementation. This change also adds two functions to "get" and "set"
13239 the implementation prior to it being automatically set the first time
13240 any other ERR function takes place. Ie. an application can call "get",
13241 pass the return value to a module it has just loaded, and that module
13242 can call its own "set" function using that value. This means the
13243 module's "ERR" operations will use (and modify) the error state in the
13244 application and not in its own statically linked copy of OpenSSL code.
13248 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13249 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13250 the operation, and provides a more encapsulated way for external code
13251 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13252 to use these functions rather than manually incrementing the counts.
13254 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13258 * Add EVP test program.
13262 * Add symmetric cipher support to ENGINE. Expect the API to change!
13266 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13267 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13268 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13269 These allow a CRL to be built without having to access X509_CRL fields
13270 directly. Modify 'ca' application to use new functions.
13274 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13275 bug workarounds. Rollback attack detection is a security feature.
13276 The problem will only arise on OpenSSL servers when TLSv1 is not
13277 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13278 Software authors not wanting to support TLSv1 will have special reasons
13279 for their choice and can explicitly enable this option.
13281 *Bodo Moeller, Lutz Jaenicke*
13283 * Rationalise EVP so it can be extended: don't include a union of
13284 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13285 (similar to those existing for EVP_CIPHER_CTX).
13290 EVP_MD_CTX_init(&md); /* new function call */
13291 EVP_DigestInit(&md, EVP_sha1());
13292 EVP_DigestUpdate(&md, in, len);
13293 EVP_DigestFinal(&md, out, NULL);
13294 EVP_MD_CTX_cleanup(&md); /* new function call */
13298 * Make DES key schedule conform to the usual scheme, as well as
13299 correcting its structure. This means that calls to DES functions
13300 now have to pass a pointer to a des_key_schedule instead of a
13301 plain des_key_schedule (which was actually always a pointer
13304 des_key_schedule ks;
13306 des_set_key_checked(..., &ks);
13307 des_ncbc_encrypt(..., &ks, ...);
13309 (Note that a later change renames 'des_...' into 'DES_...'.)
13313 * Initial reduction of linker bloat: the use of some functions, such as
13314 PEM causes large amounts of unused functions to be linked in due to
13315 poor organisation. For example pem_all.c contains every PEM function
13316 which has a knock on effect of linking in large amounts of (unused)
13317 ASN1 code. Grouping together similar functions and splitting unrelated
13318 functions prevents this.
13322 * Cleanup of EVP macros.
13326 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13327 correct `_ecb suffix`.
13331 * Add initial OCSP responder support to ocsp application. The
13332 revocation information is handled using the text based index
13333 use by the ca application. The responder can either handle
13334 requests generated internally, supplied in files (for example
13335 via a CGI script) or using an internal minimal server.
13339 * Add configuration choices to get zlib compression for TLS.
13343 * Changes to Kerberos SSL for RFC 2712 compliance:
13344 1. Implemented real KerberosWrapper, instead of just using
13345 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13346 2. Implemented optional authenticator field of KerberosWrapper.
13348 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13349 and authenticator structs; see crypto/krb5/.
13351 Generalized Kerberos calls to support multiple Kerberos libraries.
13352 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13353 via Richard Levitte*
13355 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13356 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13357 values for each of the key sizes rather than having just
13358 parameters (and 'speed' generating keys each time).
13362 * Speed up EVP routines.
13365 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13366 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13367 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13368 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13370 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13371 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13372 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13375 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13377 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13381 * Added the OS2-EMX target.
13383 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13385 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13386 New functions to support `NCONF` routines in extension code.
13387 New function `CONF_set_nconf()`
13388 to allow functions which take an `NCONF` to also handle the old `LHASH`
13389 structure: this means that the old `CONF` compatible routines can be
13390 retained (in particular w.rt. extensions) without having to duplicate the
13391 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13395 * Enhance the general user interface with mechanisms for inner control
13396 and with possibilities to have yes/no kind of prompts.
13400 * Change all calls to low-level digest routines in the library and
13401 applications to use EVP. Add missing calls to HMAC_cleanup() and
13402 don't assume HMAC_CTX can be copied using memcpy().
13404 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13406 * Add the possibility to control engines through control names but with
13407 arbitrary arguments instead of just a string.
13408 Change the key loaders to take a UI_METHOD instead of a callback
13409 function pointer. NOTE: this breaks binary compatibility with earlier
13410 versions of OpenSSL [engine].
13411 Adapt the nCipher code for these new conditions and add a card insertion
13416 * Enhance the general user interface with mechanisms to better support
13417 dialog box interfaces, application-defined prompts, the possibility
13418 to use defaults (for example default passwords from somewhere else)
13419 and interrupts/cancellations.
13423 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13424 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13428 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13429 tidy up some unnecessarily weird code in 'sk_new()').
13431 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13433 * Change the key loading routines for ENGINEs to use the same kind
13434 callback (pem_password_cb) as all other routines that need this
13439 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13440 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13441 than this minimum value is recommended.
13445 * New random seeder for OpenVMS, using the system process statistics
13446 that are easily reachable.
13450 * Windows apparently can't transparently handle global
13451 variables defined in DLLs. Initialisations such as:
13453 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13455 won't compile. This is used by the any applications that need to
13456 declare their own ASN1 modules. This was fixed by adding the option
13457 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13458 needed for static libraries under Win32.
13462 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13463 setting of purpose and trust fields. New X509_STORE trust and
13464 purpose functions and tidy up setting in other SSL functions.
13468 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13469 structure. These are inherited by X509_STORE_CTX when it is
13470 initialised. This allows various defaults to be set in the
13471 X509_STORE structure (such as flags for CRL checking and custom
13472 purpose or trust settings) for functions which only use X509_STORE_CTX
13473 internally such as S/MIME.
13475 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13476 trust settings if they are not set in X509_STORE. This allows X509_STORE
13477 purposes and trust (in S/MIME for example) to override any set by default.
13479 Add command line options for CRL checking to smime, s_client and s_server
13484 * Initial CRL based revocation checking. If the CRL checking flag(s)
13485 are set then the CRL is looked up in the X509_STORE structure and
13486 its validity and signature checked, then if the certificate is found
13487 in the CRL the verify fails with a revoked error.
13489 Various new CRL related callbacks added to X509_STORE_CTX structure.
13491 Command line options added to 'verify' application to support this.
13493 This needs some additional work, such as being able to handle multiple
13494 CRLs with different times, extension based lookup (rather than just
13495 by subject name) and ultimately more complete V2 CRL extension
13500 * Add a general user interface API (crypto/ui/). This is designed
13501 to replace things like des_read_password and friends (backward
13502 compatibility functions using this new API are provided).
13503 The purpose is to remove prompting functions from the DES code
13504 section as well as provide for prompting through dialog boxes in
13505 a window system and the like.
13509 * Add "ex_data" support to ENGINE so implementations can add state at a
13510 per-structure level rather than having to store it globally.
13514 * Make it possible for ENGINE structures to be copied when retrieved by
13515 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13516 This causes the "original" ENGINE structure to act like a template,
13517 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13518 operational state can be localised to each ENGINE structure, despite the
13519 fact they all share the same "methods". New ENGINE structures returned in
13520 this case have no functional references and the return value is the single
13521 structural reference. This matches the single structural reference returned
13522 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13527 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13528 needs to match any other type at all we need to manually clear the
13533 * Changes to the "openssl engine" utility to include;
13534 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13535 about an ENGINE's available control commands.
13536 - executing control commands from command line arguments using the
13537 '-pre' and '-post' switches. '-post' is only used if '-t' is
13538 specified and the ENGINE is successfully initialised. The syntax for
13539 the individual commands are colon-separated, for example;
13540 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13544 * New dynamic control command support for ENGINEs. ENGINEs can now
13545 declare their own commands (numbers), names (strings), descriptions,
13546 and input types for run-time discovery by calling applications. A
13547 subset of these commands are implicitly classed as "executable"
13548 depending on their input type, and only these can be invoked through
13549 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13550 can be based on user input, config files, etc). The distinction is
13551 that "executable" commands cannot return anything other than a boolean
13552 result and can only support numeric or string input, whereas some
13553 discoverable commands may only be for direct use through
13554 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13555 pointers, or other custom uses. The "executable" commands are to
13556 support parameterisations of ENGINE behaviour that can be
13557 unambiguously defined by ENGINEs and used consistently across any
13558 OpenSSL-based application. Commands have been added to all the
13559 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13560 control over shared-library paths without source code alterations.
13564 * Changed all ENGINE implementations to dynamically allocate their
13565 ENGINEs rather than declaring them statically. Apart from this being
13566 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13567 this also allows the implementations to compile without using the
13568 internal engine_int.h header.
13572 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13573 'const' value. Any code that should be able to modify a RAND_METHOD
13574 should already have non-const pointers to it (ie. they should only
13575 modify their own ones).
13579 * Made a variety of little tweaks to the ENGINE code.
13580 - "atalla" and "ubsec" string definitions were moved from header files
13581 to C code. "nuron" string definitions were placed in variables
13582 rather than hard-coded - allowing parameterisation of these values
13583 later on via ctrl() commands.
13584 - Removed unused "#if 0"'d code.
13585 - Fixed engine list iteration code so it uses ENGINE_free() to release
13586 structural references.
13587 - Constified the RAND_METHOD element of ENGINE structures.
13588 - Constified various get/set functions as appropriate and added
13589 missing functions (including a catch-all ENGINE_cpy that duplicates
13590 all ENGINE values onto a new ENGINE except reference counts/state).
13591 - Removed NULL parameter checks in get/set functions. Setting a method
13592 or function to NULL is a way of cancelling out a previously set
13593 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13594 and doesn't justify the extra error symbols and code.
13595 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13596 flags from engine_int.h to engine.h.
13597 - Changed prototypes for ENGINE handler functions (init(), finish(),
13598 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13602 * Implement binary inversion algorithm for BN_mod_inverse in addition
13603 to the algorithm using long division. The binary algorithm can be
13604 used only if the modulus is odd. On 32-bit systems, it is faster
13605 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13606 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13607 up to 450 bits. In 64-bit environments, the binary algorithm
13608 appears to be advantageous for much longer moduli; here we use it
13609 for moduli up to 2048 bits.
13613 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13614 could not support the combine flag in choice fields.
13618 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13619 extensions from a certificate request to the certificate.
13623 * Allow multiple 'certopt' and 'nameopt' options to be separated
13624 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13625 file: this allows the display of the certificate about to be
13626 signed to be customised, to allow certain fields to be included
13627 or excluded and extension details. The old system didn't display
13628 multicharacter strings properly, omitted fields not in the policy
13629 and couldn't display additional details such as extensions.
13633 * Function EC_POINTs_mul for multiple scalar multiplication
13634 of an arbitrary number of elliptic curve points
13635 \sum scalars[i]*points[i],
13636 optionally including the generator defined for the EC_GROUP:
13637 scalar*generator + \sum scalars[i]*points[i].
13639 EC_POINT_mul is a simple wrapper function for the typical case
13640 that the point list has just one item (besides the optional
13645 * First EC_METHODs for curves over GF(p):
13647 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13648 operations and provides various method functions that can also
13649 operate with faster implementations of modular arithmetic.
13651 EC_GFp_mont_method() reuses most functions that are part of
13652 EC_GFp_simple_method, but uses Montgomery arithmetic.
13654 *Bodo Moeller; point addition and point doubling
13655 implementation directly derived from source code provided by
13656 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13658 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13659 crypto/ec/ec_lib.c):
13661 Curves are EC_GROUP objects (with an optional group generator)
13662 based on EC_METHODs that are built into the library.
13664 Points are EC_POINT objects based on EC_GROUP objects.
13666 Most of the framework would be able to handle curves over arbitrary
13667 finite fields, but as there are no obvious types for fields other
13668 than GF(p), some functions are limited to that for now.
13672 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13673 that the file contains a complete HTTP response.
13677 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13678 change the def and num file printf format specifier from "%-40sXXX"
13679 to "%-39s XXX". The latter will always guarantee a space after the
13680 field while the former will cause them to run together if the field
13681 is 40 of more characters long.
13685 * Constify the cipher and digest 'method' functions and structures
13686 and modify related functions to take constant EVP_MD and EVP_CIPHER
13691 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13692 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13696 * Modify `EVP_Digest*()` routines so they now return values. Although the
13697 internal software routines can never fail additional hardware versions
13702 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13704 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13705 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13708 ERR_R_NESTED_ASN1_ERROR
13710 ERR_R_MISSING_ASN1_EOS
13711 were 4 .. 9, conflicting with
13712 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13714 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13715 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13717 Add new error code 'ERR_R_INTERNAL_ERROR'.
13721 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13726 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13727 sets the subject name for a new request or supersedes the
13728 subject name in a given request. Formats that can be parsed are
13729 'CN=Some Name, OU=myOU, C=IT'
13731 'CN=Some Name/OU=myOU/C=IT'.
13733 Add options '-batch' and '-verbose' to 'openssl req'.
13735 *Massimiliano Pala <madwolf@hackmasters.net>*
13737 * Introduce the possibility to access global variables through
13738 functions on platform were that's the best way to handle exporting
13739 global variables in shared libraries. To enable this functionality,
13740 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13741 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13742 is normally done by Configure or something similar).
13744 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13745 in the source file (foo.c) like this:
13747 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13748 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13750 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13751 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13753 OPENSSL_DECLARE_GLOBAL(int,foo);
13754 #define foo OPENSSL_GLOBAL_REF(foo)
13755 OPENSSL_DECLARE_GLOBAL(double,bar);
13756 #define bar OPENSSL_GLOBAL_REF(bar)
13758 The #defines are very important, and therefore so is including the
13759 header file everywhere where the defined globals are used.
13761 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13762 of ASN.1 items, but that structure is a bit different.
13764 The largest change is in util/mkdef.pl which has been enhanced with
13765 better and easier to understand logic to choose which symbols should
13766 go into the Windows .def files as well as a number of fixes and code
13767 cleanup (among others, algorithm keywords are now sorted
13768 lexicographically to avoid constant rewrites).
13772 * In BN_div() keep a copy of the sign of 'num' before writing the
13773 result to 'rm' because if rm==num the value will be overwritten
13774 and produce the wrong result if 'num' is negative: this caused
13775 problems with BN_mod() and BN_nnmod().
13779 * Function OCSP_request_verify(). This checks the signature on an
13780 OCSP request and verifies the signer certificate. The signer
13781 certificate is just checked for a generic purpose and OCSP request
13786 * Add OCSP_check_validity() function to check the validity of OCSP
13787 responses. OCSP responses are prepared in real time and may only
13788 be a few seconds old. Simply checking that the current time lies
13789 between thisUpdate and nextUpdate max reject otherwise valid responses
13790 caused by either OCSP responder or client clock inaccuracy. Instead
13791 we allow thisUpdate and nextUpdate to fall within a certain period of
13792 the current time. The age of the response can also optionally be
13793 checked. Two new options -validity_period and -status_age added to
13798 * If signature or public key algorithm is unrecognized print out its
13799 OID rather that just UNKNOWN.
13803 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13804 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13805 ID to be generated from the issuer certificate alone which can then be
13806 passed to OCSP_id_issuer_cmp().
13810 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13811 ASN1 modules to export functions returning ASN1_ITEM pointers
13812 instead of the ASN1_ITEM structures themselves. This adds several
13813 new macros which allow the underlying ASN1 function/structure to
13814 be accessed transparently. As a result code should not use ASN1_ITEM
13815 references directly (such as &X509_it) but instead use the relevant
13816 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13817 use of the new ASN1 code on platforms where exporting structures
13818 is problematical (for example in shared libraries) but exporting
13819 functions returning pointers to structures is not.
13823 * Add support for overriding the generation of SSL/TLS session IDs.
13824 These callbacks can be registered either in an SSL_CTX or per SSL.
13825 The purpose of this is to allow applications to control, if they wish,
13826 the arbitrary values chosen for use as session IDs, particularly as it
13827 can be useful for session caching in multiple-server environments. A
13828 command-line switch for testing this (and any client code that wishes
13829 to use such a feature) has been added to "s_server".
13831 *Geoff Thorpe, Lutz Jaenicke*
13833 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13834 of the form `#if defined(...) || defined(...) || ...` and
13835 `#if !defined(...) && !defined(...) && ...`. This also avoids
13836 the growing number of special cases it was previously handling.
13840 * Make all configuration macros available for application by making
13841 sure they are available in opensslconf.h, by giving them names starting
13842 with `OPENSSL_` to avoid conflicts with other packages and by making
13843 sure e_os2.h will cover all platform-specific cases together with
13845 Additionally, it is now possible to define configuration/platform-
13846 specific names (called "system identities"). In the C code, these
13847 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13848 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13849 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13854 * New option -set_serial to 'req' and 'x509' this allows the serial
13855 number to use to be specified on the command line. Previously self
13856 signed certificates were hard coded with serial number 0 and the
13857 CA options of 'x509' had to use a serial number in a file which was
13862 * New options to 'ca' utility to support V2 CRL entry extensions.
13863 Currently CRL reason, invalidity date and hold instruction are
13864 supported. Add new CRL extensions to V3 code and some new objects.
13868 * New function EVP_CIPHER_CTX_set_padding() this is used to
13869 disable standard block padding (aka PKCS#5 padding) in the EVP
13870 API, which was previously mandatory. This means that the data is
13871 not padded in any way and so the total length much be a multiple
13872 of the block size, otherwise an error occurs.
13876 * Initial (incomplete) OCSP SSL support.
13880 * New function OCSP_parse_url(). This splits up a URL into its host,
13881 port and path components: primarily to parse OCSP URLs. New -url
13882 option to ocsp utility.
13886 * New nonce behavior. The return value of OCSP_check_nonce() now
13887 reflects the various checks performed. Applications can decide
13888 whether to tolerate certain situations such as an absent nonce
13889 in a response when one was present in a request: the ocsp application
13890 just prints out a warning. New function OCSP_add1_basic_nonce()
13891 this is to allow responders to include a nonce in a response even if
13892 the request is nonce-less.
13896 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13897 skipped when using openssl x509 multiple times on a single input file,
13898 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13902 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13903 set string type: to handle setting ASN1_TIME structures. Fix ca
13904 utility to correctly initialize revocation date of CRLs.
13908 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13909 the clients preferred ciphersuites and rather use its own preferences.
13910 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13911 Internet Explorer by ensuring unchanged hash method during stepup.
13912 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13916 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13917 to aes and add a new 'exist' option to print out symbols that don't
13922 * Additional options to ocsp utility to allow flags to be set and
13923 additional certificates supplied.
13927 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13928 OCSP client a number of certificate to only verify the response
13933 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13934 handle the new API. Currently only ECB, CBC modes supported. Add new
13937 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13938 Encryption Standard (AES) Ciphersuites for Transport Layer
13939 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13940 not enabled by default and were not part of the "ALL" ciphersuite
13941 alias because they were not yet official; they could be
13942 explicitly requested by specifying the "AESdraft" ciphersuite
13943 group alias. In the final release of OpenSSL 0.9.7, the group
13944 alias is called "AES" and is part of "ALL".)
13946 *Ben Laurie, Steve Henson, Bodo Moeller*
13948 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13949 request to response.
13953 * Functions for OCSP responders. OCSP_request_onereq_count(),
13954 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13955 extract information from a certificate request. OCSP_response_create()
13956 creates a response and optionally adds a basic response structure.
13957 OCSP_basic_add1_status() adds a complete single response to a basic
13958 response and returns the OCSP_SINGLERESP structure just added (to allow
13959 extensions to be included for example). OCSP_basic_add1_cert() adds a
13960 certificate to a basic response and OCSP_basic_sign() signs a basic
13961 response with various flags. New helper functions ASN1_TIME_check()
13962 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13963 (converts ASN1_TIME to GeneralizedTime).
13967 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13968 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13969 structure from a certificate. X509_pubkey_digest() digests the public_key
13970 contents: this is used in various key identifiers.
13974 * Make sk_sort() tolerate a NULL argument.
13976 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13978 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13979 passed by the function are trusted implicitly. If any of them signed the
13980 response then it is assumed to be valid and is not verified.
13984 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13985 to data. This was previously part of the PKCS7 ASN1 code. This
13986 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13987 *Steve Henson, reported by Kenneth R. Robinette
13988 <support@securenetterm.com>*
13990 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13991 routines: without these tracing memory leaks is very painful.
13992 Fix leaks in PKCS12 and PKCS7 routines.
13996 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13997 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13998 effectively meant GeneralizedTime would never be used. Now it
13999 is initialised to -1 but X509_time_adj() now has to check the value
14000 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14001 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14002 *Steve Henson, reported by Kenneth R. Robinette
14003 <support@securenetterm.com>*
14005 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14006 result in a zero length in the ASN1_INTEGER structure which was
14007 not consistent with the structure when d2i_ASN1_INTEGER() was used
14008 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14009 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14010 where it did not print out a minus for negative ASN1_INTEGER.
14014 * Add summary printout to ocsp utility. The various functions which
14015 convert status values to strings have been renamed to:
14016 OCSP_response_status_str(), OCSP_cert_status_str() and
14017 OCSP_crl_reason_str() and are no longer static. New options
14018 to verify nonce values and to disable verification. OCSP response
14019 printout format cleaned up.
14023 * Add additional OCSP certificate checks. These are those specified
14024 in RFC2560. This consists of two separate checks: the CA of the
14025 certificate being checked must either be the OCSP signer certificate
14026 or the issuer of the OCSP signer certificate. In the latter case the
14027 OCSP signer certificate must contain the OCSP signing extended key
14028 usage. This check is performed by attempting to match the OCSP
14029 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14030 in the OCSP_CERTID structures of the response.
14034 * Initial OCSP certificate verification added to OCSP_basic_verify()
14035 and related routines. This uses the standard OpenSSL certificate
14036 verify routines to perform initial checks (just CA validity) and
14037 to obtain the certificate chain. Then additional checks will be
14038 performed on the chain. Currently the root CA is checked to see
14039 if it is explicitly trusted for OCSP signing. This is used to set
14040 a root CA as a global signing root: that is any certificate that
14041 chains to that CA is an acceptable OCSP signing certificate.
14045 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14046 extensions from a separate configuration file.
14047 As when reading extensions from the main configuration file,
14048 the '-extensions ...' option may be used for specifying the
14051 *Massimiliano Pala <madwolf@comune.modena.it>*
14053 * New OCSP utility. Allows OCSP requests to be generated or
14054 read. The request can be sent to a responder and the output
14055 parsed, outputted or printed in text form. Not complete yet:
14056 still needs to check the OCSP response validity.
14060 * New subcommands for 'openssl ca':
14061 `openssl ca -status <serial>` prints the status of the cert with
14062 the given serial number (according to the index file).
14063 `openssl ca -updatedb` updates the expiry status of certificates
14066 *Massimiliano Pala <madwolf@comune.modena.it>*
14068 * New '-newreq-nodes' command option to CA.pl. This is like
14069 '-newreq', but calls 'openssl req' with the '-nodes' option
14070 so that the resulting key is not encrypted.
14072 *Damien Miller <djm@mindrot.org>*
14074 * New configuration for the GNU Hurd.
14076 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14078 * Initial code to implement OCSP basic response verify. This
14079 is currently incomplete. Currently just finds the signer's
14080 certificate and verifies the signature on the response.
14084 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14085 value of OPENSSLDIR. This is available via the new '-d' option
14086 to 'openssl version', and is also included in 'openssl version -a'.
14090 * Allowing defining memory allocation callbacks that will be given
14091 file name and line number information in additional arguments
14092 (a `const char*` and an int). The basic functionality remains, as
14093 well as the original possibility to just replace malloc(),
14094 realloc() and free() by functions that do not know about these
14095 additional arguments. To register and find out the current
14096 settings for extended allocation functions, the following
14097 functions are provided:
14099 CRYPTO_set_mem_ex_functions
14100 CRYPTO_set_locked_mem_ex_functions
14101 CRYPTO_get_mem_ex_functions
14102 CRYPTO_get_locked_mem_ex_functions
14104 These work the same way as CRYPTO_set_mem_functions and friends.
14105 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14106 extended allocation function is enabled.
14107 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14108 a conventional allocation function is enabled.
14110 *Richard Levitte, Bodo Moeller*
14112 * Finish off removing the remaining LHASH function pointer casts.
14113 There should no longer be any prototype-casting required when using
14114 the LHASH abstraction, and any casts that remain are "bugs". See
14115 the callback types and macros at the head of lhash.h for details
14116 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14120 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14121 If /dev/[u]random devices are not available or do not return enough
14122 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14124 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14125 /etc/entropy will be queried once each in this sequence, querying stops
14126 when enough entropy was collected without querying more sockets.
14130 * Change the Unix RAND_poll() variant to be able to poll several
14131 random devices, as specified by DEVRANDOM, until a sufficient amount
14132 of data has been collected. We spend at most 10 ms on each file
14133 (select timeout) and read in non-blocking mode. DEVRANDOM now
14134 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14135 (previously it was just the string "/dev/urandom"), so on typical
14136 platforms the 10 ms delay will never occur.
14137 Also separate out the Unix variant to its own file, rand_unix.c.
14138 For VMS, there's a currently-empty rand_vms.c.
14142 * Move OCSP client related routines to ocsp_cl.c. These
14143 provide utility functions which an application needing
14144 to issue a request to an OCSP responder and analyse the
14145 response will typically need: as opposed to those which an
14146 OCSP responder itself would need which will be added later.
14148 OCSP_request_sign() signs an OCSP request with an API similar
14149 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14150 response. OCSP_response_get1_basic() extracts basic response
14151 from response. OCSP_resp_find_status(): finds and extracts status
14152 information from an OCSP_CERTID structure (which will be created
14153 when the request structure is built). These are built from lower
14154 level functions which work on OCSP_SINGLERESP structures but
14155 won't normally be used unless the application wishes to examine
14156 extensions in the OCSP response for example.
14158 Replace nonce routines with a pair of functions.
14159 OCSP_request_add1_nonce() adds a nonce value and optionally
14160 generates a random value. OCSP_check_nonce() checks the
14161 validity of the nonce in an OCSP response.
14165 * Change function OCSP_request_add() to OCSP_request_add0_id().
14166 This doesn't copy the supplied OCSP_CERTID and avoids the
14167 need to free up the newly created id. Change return type
14168 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14169 This can then be used to add extensions to the request.
14170 Deleted OCSP_request_new(), since most of its functionality
14171 is now in OCSP_REQUEST_new() (and the case insensitive name
14172 clash) apart from the ability to set the request name which
14173 will be added elsewhere.
14177 * Update OCSP API. Remove obsolete extensions argument from
14178 various functions. Extensions are now handled using the new
14179 OCSP extension code. New simple OCSP HTTP function which
14180 can be used to send requests and parse the response.
14184 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14185 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14186 uses the special reorder version of SET OF to sort the attributes
14187 and reorder them to match the encoded order. This resolves a long
14188 standing problem: a verify on a PKCS7 structure just after signing
14189 it used to fail because the attribute order did not match the
14190 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14191 it uses the received order. This is necessary to tolerate some broken
14192 software that does not order SET OF. This is handled by encoding
14193 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14194 to produce the required SET OF.
14198 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14199 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14200 files to get correct declarations of the ASN.1 item variables.
14204 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14205 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14206 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14207 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14208 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14209 ASN1_ITEM and no wrapper functions.
14213 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14214 replace the old function pointer based I/O routines. Change most of
14215 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14219 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14220 lines, recognize more "algorithms" that can be deselected, and make
14221 it complain about algorithm deselection that isn't recognised.
14225 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14226 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14227 to use new functions. Add NO_ASN1_OLD which can be set to remove
14228 some old style ASN1 functions: this can be used to determine if old
14229 code will still work when these eventually go away.
14233 * New extension functions for OCSP structures, these follow the
14234 same conventions as certificates and CRLs.
14238 * New function X509V3_add1_i2d(). This automatically encodes and
14239 adds an extension. Its behaviour can be customised with various
14240 flags to append, replace or delete. Various wrappers added for
14241 certificates and CRLs.
14245 * Fix to avoid calling the underlying ASN1 print routine when
14246 an extension cannot be parsed. Correct a typo in the
14247 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14251 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14252 entries for variables.
14256 * Add functionality to `apps/openssl.c` for detecting locking
14257 problems: As the program is single-threaded, all we have
14258 to do is register a locking callback using an array for
14259 storing which locks are currently held by the program.
14263 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14264 SSL_get_ex_data_X509_STORE_idx(), which is used in
14265 ssl_verify_cert_chain() and thus can be called at any time
14266 during TLS/SSL handshakes so that thread-safety is essential.
14267 Unfortunately, the ex_data design is not at all suited
14268 for multi-threaded use, so it probably should be abolished.
14272 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14274 *Broadcom, tweaked and integrated by Geoff Thorpe*
14276 * Move common extension printing code to new function
14277 X509V3_print_extensions(). Reorganise OCSP print routines and
14278 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14282 * New function X509_signature_print() to remove duplication in some
14287 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14288 set (this was treated exactly the same as SET OF previously). This
14289 is used to reorder the STACK representing the structure to match the
14290 encoding. This will be used to get round a problem where a PKCS7
14291 structure which was signed could not be verified because the STACK
14292 order did not reflect the encoded order.
14296 * Reimplement the OCSP ASN1 module using the new code.
14300 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14301 for its ASN1 operations. The old style function pointers still exist
14302 for now but they will eventually go away.
14306 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14307 completely replaces the old ASN1 functionality with a table driven
14308 encoder and decoder which interprets an ASN1_ITEM structure describing
14309 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14310 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14311 has also been converted to the new form.
14315 * Change BN_mod_exp_recp so that negative moduli are tolerated
14316 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14317 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14318 for negative moduli.
14322 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14323 of not touching the result's sign bit.
14327 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14332 * Changed the LHASH code to use prototypes for callbacks, and created
14333 macros to declare and implement thin (optionally static) functions
14334 that provide type-safety and avoid function pointer casting for the
14335 type-specific callbacks.
14339 * Added Kerberos Cipher Suites to be used with TLS, as written in
14341 *Veers Staats <staatsvr@asc.hpc.mil>,
14342 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14344 * Reformat the FAQ so the different questions and answers can be divided
14345 in sections depending on the subject.
14349 * Have the zlib compression code load ZLIB.DLL dynamically under
14354 * New function BN_mod_sqrt for computing square roots modulo a prime
14355 (using the probabilistic Tonelli-Shanks algorithm unless
14356 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14357 be handled deterministically).
14359 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14361 * Make BN_mod_inverse faster by explicitly handling small quotients
14362 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14363 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14367 * New function BN_kronecker.
14371 * Fix BN_gcd so that it works on negative inputs; the result is
14372 positive unless both parameters are zero.
14373 Previously something reasonably close to an infinite loop was
14374 possible because numbers could be growing instead of shrinking
14375 in the implementation of Euclid's algorithm.
14379 * Fix BN_is_word() and BN_is_one() macros to take into account the
14380 sign of the number in question.
14382 Fix BN_is_word(a,w) to work correctly for w == 0.
14384 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14385 because its test if the absolute value of 'a' equals 'w'.
14386 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14387 it exists mostly for use in the implementations of BN_is_zero(),
14388 BN_is_one(), and BN_is_word().
14392 * New function BN_swap.
14396 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14397 the exponentiation functions are more likely to produce reasonable
14398 results on negative inputs.
14402 * Change BN_mod_mul so that the result is always non-negative.
14403 Previously, it could be negative if one of the factors was negative;
14404 I don't think anyone really wanted that behaviour.
14408 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14409 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14410 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14411 and add new functions:
14420 BN_mod_lshift1_quick
14422 BN_mod_lshift_quick
14424 These functions always generate non-negative results.
14426 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14427 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14429 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14430 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14431 be reduced modulo `m`.
14433 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14436 The following entry accidentally appeared in the CHANGES file
14437 distributed with OpenSSL 0.9.7. The modifications described in
14438 it do *not* apply to OpenSSL 0.9.7.
14440 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14441 was actually never needed) and in BN_mul(). The removal in BN_mul()
14442 required a small change in bn_mul_part_recursive() and the addition
14443 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14444 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14445 bn_sub_words() and bn_add_words() except they take arrays with
14451 * In 'openssl passwd', verify passwords read from the terminal
14452 unless the '-salt' option is used (which usually means that
14453 verification would just waste user's time since the resulting
14454 hash is going to be compared with some given password hash)
14455 or the new '-noverify' option is used.
14457 This is an incompatible change, but it does not affect
14458 non-interactive use of 'openssl passwd' (passwords on the command
14459 line, '-stdin' option, '-in ...' option) and thus should not
14460 cause any problems.
14464 * Remove all references to RSAref, since there's no more need for it.
14468 * Make DSO load along a path given through an environment variable
14469 (SHLIB_PATH) with shl_load().
14473 * Constify the ENGINE code as a result of BIGNUM constification.
14474 Also constify the RSA code and most things related to it. In a
14475 few places, most notable in the depth of the ASN.1 code, ugly
14476 casts back to non-const were required (to be solved at a later
14481 * Make it so the openssl application has all engines loaded by default.
14485 * Constify the BIGNUM routines a little more.
14489 * Add the following functions:
14491 ENGINE_load_cswift()
14493 ENGINE_load_atalla()
14494 ENGINE_load_nuron()
14495 ENGINE_load_builtin_engines()
14497 That way, an application can itself choose if external engines that
14498 are built-in in OpenSSL shall ever be used or not. The benefit is
14499 that applications won't have to be linked with libdl or other dso
14500 libraries unless it's really needed.
14502 Changed 'openssl engine' to load all engines on demand.
14503 Changed the engine header files to avoid the duplication of some
14504 declarations (they differed!).
14508 * 'openssl engine' can now list capabilities.
14512 * Better error reporting in 'openssl engine'.
14516 * Never call load_dh_param(NULL) in s_server.
14520 * Add engine application. It can currently list engines by name and
14521 identity, and test if they are actually available.
14525 * Improve RPM specification file by forcing symbolic linking and making
14526 sure the installed documentation is also owned by root.root.
14528 *Damien Miller <djm@mindrot.org>*
14530 * Give the OpenSSL applications more possibilities to make use of
14531 keys (public as well as private) handled by engines.
14535 * Add OCSP code that comes from CertCo.
14539 * Add VMS support for the Rijndael code.
14543 * Added untested support for Nuron crypto accelerator.
14547 * Add support for external cryptographic devices. This code was
14548 previously distributed separately as the "engine" branch.
14550 *Geoff Thorpe, Richard Levitte*
14552 * Rework the filename-translation in the DSO code. It is now possible to
14553 have far greater control over how a "name" is turned into a filename
14554 depending on the operating environment and any oddities about the
14555 different shared library filenames on each system.
14559 * Support threads on FreeBSD-elf in Configure.
14563 * Fix for SHA1 assembly problem with MASM: it produces
14564 warnings about corrupt line number information when assembling
14565 with debugging information. This is caused by the overlapping
14568 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14571 NCONF_get_number() has no error checking at all. As a replacement,
14572 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14573 promoted strongly. The old NCONF_get_number is kept around for
14574 binary backward compatibility.
14575 Make it possible for methods to load from something other than a BIO,
14576 by providing a function pointer that is given a name instead of a BIO.
14577 For example, this could be used to load configuration data from an
14582 * Fix for non blocking accept BIOs. Added new I/O special reason
14583 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14584 with non blocking I/O was not possible because no retry code was
14585 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14590 * Added the beginnings of Rijndael support.
14594 * Fix for bug in DirectoryString mask setting. Add support for
14595 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14596 to allow certificate printing to more controllable, additional
14597 'certopt' option to 'x509' to allow new printing options to be
14602 * Clean old EAY MD5 hack from e_os.h.
14606 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14608 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14609 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14611 *Joe Orton, Steve Henson*
14613 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14615 * Fix additional bug revealed by the NISCC test suite:
14617 Stop bug triggering large recursion when presented with
14618 certain ASN.1 tags ([CVE-2003-0851])
14622 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14624 * Fix various bugs revealed by running the NISCC test suite:
14626 Stop out of bounds reads in the ASN1 code when presented with
14627 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14629 If verify callback ignores invalid public key errors don't try to check
14630 certificate signature with the NULL public key.
14634 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14635 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14640 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14641 extra data after the compression methods not only for TLS 1.0
14642 but also for SSL 3.0 (as required by the specification).
14644 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14646 * Change X509_certificate_type() to mark the key as exported/exportable
14647 when it's 512 *bits* long, not 512 bytes.
14651 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14653 * Countermeasure against the Klima-Pokorny-Rosa extension of
14654 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14655 a protocol version number mismatch like a decryption error
14656 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14660 * Turn on RSA blinding by default in the default implementation
14661 to avoid a timing attack. Applications that don't want it can call
14662 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14663 They would be ill-advised to do so in most cases.
14665 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14667 * Change RSA blinding code so that it works when the PRNG is not
14668 seeded (in this case, the secret RSA exponent is abused as
14669 an unpredictable seed -- if it is not unpredictable, there
14670 is no point in blinding anyway). Make RSA blinding thread-safe
14671 by remembering the creator's thread ID in rsa->blinding and
14672 having all other threads use local one-time blinding factors
14673 (this requires more computation than sharing rsa->blinding, but
14674 avoids excessive locking; and if an RSA object is not shared
14675 between threads, blinding will still be very fast).
14679 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14681 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14682 via timing by performing a MAC computation even if incorrect
14683 block cipher padding has been found. This is a countermeasure
14684 against active attacks where the attacker has to distinguish
14685 between bad padding and a MAC verification error. ([CVE-2003-0078])
14687 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14688 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14689 Martin Vuagnoux (EPFL, Ilion)*
14691 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14693 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14694 memory from its contents. This is done with a counter that will
14695 place alternating values in each byte. This can be used to solve
14696 two issues: 1) the removal of calls to memset() by highly optimizing
14697 compilers, and 2) cleansing with other values than 0, since those can
14698 be read through on certain media, for example a swap space on disk.
14702 * Bugfix: client side session caching did not work with external caching,
14703 because the session->cipher setting was not restored when reloading
14704 from the external cache. This problem was masked, when
14705 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14706 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14710 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14711 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14713 *Zeev Lieber <zeev-l@yahoo.com>*
14715 * Undo an undocumented change introduced in 0.9.6e which caused
14716 repeated calls to OpenSSL_add_all_ciphers() and
14717 OpenSSL_add_all_digests() to be ignored, even after calling
14722 * Change the default configuration reader to deal with last line not
14723 being properly terminated.
14727 * Change X509_NAME_cmp() so it applies the special rules on handling
14728 DN values that are of type PrintableString, as well as RDNs of type
14729 emailAddress where the value has the type ia5String.
14731 *stefank@valicert.com via Richard Levitte*
14733 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14734 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14735 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14736 the bitwise-OR of the two for use by the majority of applications
14737 wanting this behaviour, and update the docs. The documented
14738 behaviour and actual behaviour were inconsistent and had been
14739 changing anyway, so this is more a bug-fix than a behavioural
14742 *Geoff Thorpe, diagnosed by Nadav Har'El*
14744 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14745 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14749 * Fix initialization code race conditions in
14750 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14751 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14752 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14753 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14754 ssl2_get_cipher_by_char(),
14755 ssl3_get_cipher_by_char().
14757 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14759 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14760 the cached sessions are flushed, as the remove_cb() might use ex_data
14761 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14762 (see [openssl.org #212]).
14764 *Geoff Thorpe, Lutz Jaenicke*
14766 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14767 length, instead of the encoding length to d2i_ASN1_OBJECT.
14771 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14773 * [In 0.9.6g-engine release:]
14774 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14776 *Lynn Gazis <lgazis@rainbow.com>*
14778 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14780 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14781 and get fix the header length calculation.
14782 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14783 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14785 * Use proper error handling instead of 'assertions' in buffer
14786 overflow checks added in 0.9.6e. This prevents DoS (the
14787 assertions could call abort()).
14789 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14791 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14793 * Add various sanity checks to asn1_get_length() to reject
14794 the ASN1 length bytes if they exceed sizeof(long), will appear
14795 negative or the content length exceeds the length of the
14798 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14800 * Fix cipher selection routines: ciphers without encryption had no flags
14801 for the cipher strength set and where therefore not handled correctly
14802 by the selection routines (PR #130).
14806 * Fix EVP_dsa_sha macro.
14811 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14812 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14813 that was added in OpenSSL 0.9.6d.
14815 As the countermeasure turned out to be incompatible with some
14816 broken SSL implementations, the new option is part of SSL_OP_ALL.
14817 SSL_OP_ALL is usually employed when compatibility with weird SSL
14818 implementations is desired (e.g. '-bugs' option to 's_client' and
14819 's_server'), so the new option is automatically set in many
14824 * Changes in security patch:
14826 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14827 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14828 Air Force Materiel Command, USAF, under agreement number
14831 * Add various sanity checks to asn1_get_length() to reject
14832 the ASN1 length bytes if they exceed sizeof(long), will appear
14833 negative or the content length exceeds the length of the
14834 supplied buffer. ([CVE-2002-0659])
14836 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14838 * Assertions for various potential buffer overflows, not known to
14839 happen in practice.
14841 *Ben Laurie (CHATS)*
14843 * Various temporary buffers to hold ASCII versions of integers were
14844 too small for 64 bit platforms. ([CVE-2002-0655])
14845 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14847 * Remote buffer overflow in SSL3 protocol - an attacker could
14848 supply an oversized session ID to a client. ([CVE-2002-0656])
14850 *Ben Laurie (CHATS)*
14852 * Remote buffer overflow in SSL2 protocol - an attacker could
14853 supply an oversized client master key. ([CVE-2002-0656])
14855 *Ben Laurie (CHATS)*
14857 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14859 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14860 encoded as NULL) with id-dsa-with-sha1.
14862 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14864 * Check various `X509_...()` return values in `apps/req.c`.
14866 *Nils Larsch <nla@trustcenter.de>*
14868 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14869 an end-of-file condition would erroneously be flagged, when the CRLF
14870 was just at the end of a processed block. The bug was discovered when
14871 processing data through a buffering memory BIO handing the data to a
14872 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14873 <ptsekov@syntrex.com> and Nedelcho Stanev.
14877 * Implement a countermeasure against a vulnerability recently found
14878 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14879 before application data chunks to avoid the use of known IVs
14880 with data potentially chosen by the attacker.
14884 * Fix length checks in ssl3_get_client_hello().
14888 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14889 to prevent ssl3_read_internal() from incorrectly assuming that
14890 ssl3_read_bytes() found application data while handshake
14891 processing was enabled when in fact s->s3->in_read_app_data was
14892 merely automatically cleared during the initial handshake.
14894 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14896 * Fix object definitions for Private and Enterprise: they were not
14897 recognized in their shortname (=lowercase) representation. Extend
14898 obj_dat.pl to issue an error when using undefined keywords instead
14899 of silently ignoring the problem (Svenning Sorensen
14900 <sss@sss.dnsalias.net>).
14904 * Fix DH_generate_parameters() so that it works for 'non-standard'
14905 generators, i.e. generators other than 2 and 5. (Previously, the
14906 code did not properly initialise the 'add' and 'rem' values to
14907 BN_generate_prime().)
14909 In the new general case, we do not insist that 'generator' is
14910 actually a primitive root: This requirement is rather pointless;
14911 a generator of the order-q subgroup is just as good, if not
14916 * Map new X509 verification errors to alerts. Discovered and submitted by
14917 Tom Wu <tom@arcot.com>.
14921 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14922 returning non-zero before the data has been completely received
14923 when using non-blocking I/O.
14925 *Bodo Moeller; problem pointed out by John Hughes*
14927 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14929 *Ben Laurie, Lutz Jaenicke*
14931 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14932 Yoram Zahavi <YoramZ@gilian.com>).
14936 * Add information about CygWin 1.3 and on, and preserve proper
14937 configuration for the versions before that.
14939 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14941 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14942 check whether we deal with a copy of a session and do not delete from
14943 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14944 <izhar@checkpoint.com>.
14948 * Do not store session data into the internal session cache, if it
14949 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14950 flag is set). Proposed by Aslam <aslam@funk.com>.
14954 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14959 * [In 0.9.6d-engine release:]
14960 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14962 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14964 * Add the configuration target linux-s390x.
14966 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14968 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14969 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14970 variable as an indication that a ClientHello message has been
14971 received. As the flag value will be lost between multiple
14972 invocations of ssl3_accept when using non-blocking I/O, the
14973 function may not be aware that a handshake has actually taken
14974 place, thus preventing a new session from being added to the
14977 To avoid this problem, we now set s->new_session to 2 instead of
14978 using a local variable.
14980 *Lutz Jaenicke, Bodo Moeller*
14982 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14983 if the SSL_R_LENGTH_MISMATCH error is detected.
14985 *Geoff Thorpe, Bodo Moeller*
14987 * New 'shared_ldflag' column in Configure platform table.
14991 * Fix EVP_CIPHER_mode macro.
14993 *"Dan S. Camper" <dan@bti.net>*
14995 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14996 type, we must throw them away by setting rr->length to 0.
14998 *D P Chang <dpc@qualys.com>*
15000 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15002 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15003 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15004 worked incorrectly for those cases where range = `10..._2` and
15005 `3*range` is two bits longer than range.)
15009 * Only add signing time to PKCS7 structures if it is not already
15014 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15015 OBJ_ld_ce should be OBJ_id_ce.
15016 Also some ip-pda OIDs in crypto/objects/objects.txt were
15017 incorrect (cf. RFC 3039).
15019 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15021 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15022 returns early because it has nothing to do.
15024 *Andy Schneider <andy.schneider@bjss.co.uk>*
15026 * [In 0.9.6c-engine release:]
15027 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15029 *Andy Schneider <andy.schneider@bjss.co.uk>*
15031 * [In 0.9.6c-engine release:]
15032 Add support for Cryptographic Appliance's keyserver technology.
15033 (Use engine 'keyclient')
15035 *Cryptographic Appliances and Geoff Thorpe*
15037 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15038 is called via tools/c89.sh because arguments have to be
15039 rearranged (all '-L' options must appear before the first object
15042 *Richard Shapiro <rshapiro@abinitio.com>*
15044 * [In 0.9.6c-engine release:]
15045 Add support for Broadcom crypto accelerator cards, backported
15048 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15050 * [In 0.9.6c-engine release:]
15051 Add support for SureWare crypto accelerator cards from
15052 Baltimore Technologies. (Use engine 'sureware')
15054 *Baltimore Technologies and Mark Cox*
15056 * [In 0.9.6c-engine release:]
15057 Add support for crypto accelerator cards from Accelerated
15058 Encryption Processing, www.aep.ie. (Use engine 'aep')
15060 *AEP Inc. and Mark Cox*
15062 * Add a configuration entry for gcc on UnixWare.
15064 *Gary Benson <gbenson@redhat.com>*
15066 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15067 messages are stored in a single piece (fixed-length part and
15068 variable-length part combined) and fix various bugs found on the way.
15072 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15073 instead. BIO_gethostbyname() does not know what timeouts are
15074 appropriate, so entries would stay in cache even when they have
15076 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15078 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15079 faced with a pathologically small ClientHello fragment that does
15080 not contain client_version: Instead of aborting with an error,
15081 simply choose the highest available protocol version (i.e.,
15082 TLS 1.0 unless it is disabled). In practice, ClientHello
15083 messages are never sent like this, but this change gives us
15084 strictly correct behaviour at least for TLS.
15088 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15089 never resets s->method to s->ctx->method when called from within
15090 one of the SSL handshake functions.
15092 *Bodo Moeller; problem pointed out by Niko Baric*
15094 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15095 (sent using the client's version number) if client_version is
15096 smaller than the protocol version in use. Also change
15097 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15098 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15099 the client will at least see that alert.
15103 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15108 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15109 client receives HelloRequest while in a handshake.
15111 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15113 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15114 should end in 'break', not 'goto end' which circumvents various
15115 cleanups done in state SSL_ST_OK. But session related stuff
15116 must be disabled for SSL_ST_OK in the case that we just sent a
15119 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15120 before just sending a HelloRequest.
15122 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15124 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15125 reveal whether illegal block cipher padding was found or a MAC
15126 verification error occurred. (Neither SSLerr() codes nor alerts
15127 are directly visible to potential attackers, but the information
15128 may leak via logfiles.)
15130 Similar changes are not required for the SSL 2.0 implementation
15131 because the number of padding bytes is sent in clear for SSL 2.0,
15132 and the extra bytes are just ignored. However ssl/s2_pkt.c
15133 failed to verify that the purported number of padding bytes is in
15138 * Add OpenUNIX-8 support including shared libraries
15139 (Boyd Lynn Gerber <gerberb@zenez.com>).
15143 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15144 'wristwatch attack' using huge encoding parameters (cf.
15145 James H. Manger's CRYPTO 2001 paper). Note that the
15146 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15147 encoding parameters and hence was not vulnerable.
15151 * BN_sqr() bug fix.
15153 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15155 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15156 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15157 followed by modular reduction.
15159 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15161 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15162 equivalent based on BN_pseudo_rand() instead of BN_rand().
15166 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15167 This function was broken, as the check for a new client hello message
15168 to handle SGC did not allow these large messages.
15169 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15173 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15177 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15178 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15182 * Rework the configuration and shared library support for Tru64 Unix.
15183 The configuration part makes use of modern compiler features and
15184 still retains old compiler behavior for those that run older versions
15185 of the OS. The shared library support part includes a variant that
15186 uses the RPATH feature, and is available through the special
15187 configuration target "alpha-cc-rpath", which will never be selected
15190 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15192 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15193 with the same message size as in ssl3_get_certificate_request().
15194 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15195 messages might inadvertently be reject as too long.
15197 *Petr Lampa <lampa@fee.vutbr.cz>*
15199 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15203 * Modified SSL library such that the verify_callback that has been set
15204 specifically for an SSL object with SSL_set_verify() is actually being
15205 used. Before the change, a verify_callback set with this function was
15206 ignored and the verify_callback() set in the SSL_CTX at the time of
15207 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15208 to allow the necessary settings.
15212 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15213 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15214 done automatically (in contradiction to the requirements of the C
15215 standard). This made problems when used from OpenSSH.
15219 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15220 dh->length and always used
15222 BN_rand_range(priv_key, dh->p).
15224 BN_rand_range() is not necessary for Diffie-Hellman, and this
15225 specific range makes Diffie-Hellman unnecessarily inefficient if
15226 dh->length (recommended exponent length) is much smaller than the
15227 length of dh->p. We could use BN_rand_range() if the order of
15228 the subgroup was stored in the DH structure, but we only have
15233 BN_rand(priv_key, l, ...)
15235 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15242 RSA_eay_public_encrypt
15243 RSA_eay_private_decrypt
15244 RSA_eay_private_encrypt (signing)
15245 RSA_eay_public_decrypt (signature verification)
15247 (default implementations for RSA_public_encrypt,
15248 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15249 always reject numbers >= n.
15253 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15254 to synchronize access to 'locking_thread'. This is necessary on
15255 systems where access to 'locking_thread' (an 'unsigned long'
15256 variable) is not atomic.
15260 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15261 *before* setting the 'crypto_lock_rand' flag. The previous code had
15262 a race condition if 0 is a valid thread ID.
15264 *Travis Vitek <vitek@roguewave.com>*
15266 * Add support for shared libraries under Irix.
15268 *Albert Chin-A-Young <china@thewrittenword.com>*
15270 * Add configuration option to build on Linux on both big-endian and
15271 little-endian MIPS.
15273 *Ralf Baechle <ralf@uni-koblenz.de>*
15275 * Add the possibility to create shared libraries on HP-UX.
15279 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15281 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15282 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15283 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15284 PRNG state recovery was possible based on the output of
15285 one PRNG request appropriately sized to gain knowledge on
15286 'md' followed by enough consecutive 1-byte PRNG requests
15287 to traverse all of 'state'.
15289 1. When updating 'md_local' (the current thread's copy of 'md')
15290 during PRNG output generation, hash all of the previous
15291 'md_local' value, not just the half used for PRNG output.
15293 2. Make the number of bytes from 'state' included into the hash
15294 independent from the number of PRNG bytes requested.
15296 The first measure alone would be sufficient to avoid
15297 Markku-Juhani's attack. (Actually it had never occurred
15298 to me that the half of 'md_local' used for chaining was the
15299 half from which PRNG output bytes were taken -- I had always
15300 assumed that the secret half would be used.) The second
15301 measure makes sure that additional data from 'state' is never
15302 mixed into 'md_local' in small portions; this heuristically
15303 further strengthens the PRNG.
15307 * Fix crypto/bn/asm/mips3.s.
15311 * When only the key is given to "enc", the IV is undefined. Print out
15312 an error message in this case.
15316 * Handle special case when X509_NAME is empty in X509 printing routines.
15320 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15321 positive and less than q.
15325 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15326 used: it isn't thread safe and the add_lock_callback should handle
15329 *Paul Rose <Paul.Rose@bridge.com>*
15331 * Verify that incoming data obeys the block size in
15332 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15338 *Ulf Möller, Bodo Möller*
15340 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15341 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15342 when fixing the server behaviour for backwards-compatible 'client
15343 hello' messages. (Note that the attack is impractical against
15344 SSL 3.0 and TLS 1.0 anyway because length and version checking
15345 means that the probability of guessing a valid ciphertext is
15346 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15349 Before 0.9.5, the countermeasure (hide the error by generating a
15350 random 'decryption result') did not work properly because
15351 ERR_clear_error() was missing, meaning that SSL_get_error() would
15352 detect the supposedly ignored error.
15354 Both problems are now fixed.
15358 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15359 (previously it was 1024).
15363 * Fix for compatibility mode trust settings: ignore trust settings
15364 unless some valid trust or reject settings are present.
15368 * Fix for blowfish EVP: its a variable length cipher.
15372 * Fix various bugs related to DSA S/MIME verification. Handle missing
15373 parameters in DSA public key structures and return an error in the
15374 DSA routines if parameters are absent.
15378 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15379 in the current directory if neither $RANDFILE nor $HOME was set.
15380 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15381 caused some confusion to Windows users who haven't defined $HOME.
15382 Thus RAND_file_name() is changed again: e_os.h can define a
15383 DEFAULT_HOME, which will be used if $HOME is not set.
15384 For Windows, we use "C:"; on other platforms, we still require
15385 environment variables.
15387 * Move 'if (!initialized) RAND_poll()' into regions protected by
15388 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15389 having multiple threads call RAND_poll() concurrently.
15393 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15394 combination of a flag and a thread ID variable.
15395 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15396 flag), *other* threads can enter ssleay_add_bytes without obeying
15397 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15398 that they do not hold after the first thread unsets add_do_not_lock).
15402 * Change bctest again: '-x' expressions are not available in all
15403 versions of 'test'.
15407 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15409 * Fix a couple of memory leaks in PKCS7_dataDecode()
15411 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15413 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15414 the default extension for executables, if any. Also, make the perl
15415 scripts that use symlink() to test if it really exists and use "cp"
15416 if it doesn't. All this made OpenSSL compilable and installable in
15421 * Fix for asn1_GetSequence() for indefinite length constructed data.
15422 If SEQUENCE is length is indefinite just set c->slen to the total
15423 amount of data available.
15425 *Steve Henson, reported by shige@FreeBSD.org*
15427 *This change does not apply to 0.9.7.*
15429 * Change bctest to avoid here-documents inside command substitution
15430 (workaround for FreeBSD /bin/sh bug).
15431 For compatibility with Ultrix, avoid shell functions (introduced
15432 in the bctest version that searches along $PATH).
15436 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15437 with des_encrypt() defined on some operating systems, like Solaris
15442 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15443 On the Importance of Eliminating Errors in Cryptographic
15444 Computations, J. Cryptology 14 (2001) 2, 101-119,
15445 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15449 * MIPS assembler BIGNUM division bug fix.
15453 * Disabled incorrect Alpha assembler code.
15457 * Fix PKCS#7 decode routines so they correctly update the length
15458 after reading an EOC for the EXPLICIT tag.
15462 *This change does not apply to 0.9.7.*
15464 * Fix bug in PKCS#12 key generation routines. This was triggered
15465 if a 3DES key was generated with a 0 initial byte. Include
15466 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15467 (but broken) behaviour.
15471 * Enhance bctest to search for a working bc along $PATH and print
15474 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15476 * Fix memory leaks in err.c: free err_data string if necessary;
15477 don't write to the wrong index in ERR_set_error_data.
15481 * Implement ssl23_peek (analogous to ssl23_read), which previously
15486 * Replace rdtsc with `_emit` statements for VC++ version 5.
15488 *Jeremy Cooper <jeremy@baymoo.org>*
15490 * Make it possible to reuse SSLv2 sessions.
15494 * In copy_email() check for >= 0 as a return value for
15495 X509_NAME_get_index_by_NID() since 0 is a valid index.
15497 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15499 * Avoid coredump with unsupported or invalid public keys by checking if
15500 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15501 PKCS7_verify() fails with non detached data.
15505 * Don't use getenv in library functions when run as setuid/setgid.
15506 New function OPENSSL_issetugid().
15510 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15511 due to incorrect handling of multi-threading:
15513 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15515 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15517 3. Count how many times MemCheck_off() has been called so that
15518 nested use can be treated correctly. This also avoids
15519 inband-signalling in the previous code (which relied on the
15520 assumption that thread ID 0 is impossible).
15524 * Add "-rand" option also to s_client and s_server.
15528 * Fix CPU detection on Irix 6.x.
15529 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15530 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15532 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15537 *This change does not apply to 0.9.7.*
15539 * Use the cached encoding of an X509_NAME structure rather than
15540 copying it. This is apparently the reason for the libsafe "errors"
15541 but the code is actually correct.
15545 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15546 Bleichenbacher's DSA attack.
15547 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15548 to be set and top=0 forces the highest bit to be set; top=-1 is new
15549 and leaves the highest bit random.
15551 *Ulf Moeller, Bodo Moeller*
15553 * In the `NCONF_...`-based implementations for `CONF_...` queries
15554 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15555 a temporary CONF structure with the data component set to NULL
15556 (which gives segmentation faults in lh_retrieve).
15557 Instead, use NULL for the CONF pointer in CONF_get_string and
15558 CONF_get_number (which may use environment variables) and directly
15559 return NULL from CONF_get_section.
15563 * Fix potential buffer overrun for EBCDIC.
15567 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15568 keyUsage if basicConstraints absent for a CA.
15572 * Make SMIME_write_PKCS7() write mail header values with a format that
15573 is more generally accepted (no spaces before the semicolon), since
15574 some programs can't parse those values properly otherwise. Also make
15575 sure BIO's that break lines after each write do not create invalid
15580 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15581 macros previously used would not encode an empty SEQUENCE OF
15582 and break the signature.
15586 *This change does not apply to 0.9.7.*
15588 * Zero the premaster secret after deriving the master secret in
15593 * Add some EVP_add_digest_alias registrations (as found in
15594 OpenSSL_add_all_digests()) to SSL_library_init()
15595 aka OpenSSL_add_ssl_algorithms(). This provides improved
15596 compatibility with peers using X.509 certificates
15597 with unconventional AlgorithmIdentifier OIDs.
15601 * Fix for Irix with NO_ASM.
15603 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15605 * ./config script fixes.
15607 *Ulf Moeller, Richard Levitte*
15609 * Fix 'openssl passwd -1'.
15613 * Change PKCS12_key_gen_asc() so it can cope with non null
15614 terminated strings whose length is passed in the passlen
15615 parameter, for example from PEM callbacks. This was done
15616 by adding an extra length parameter to asc2uni().
15618 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15620 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15621 call failed, free the DSA structure.
15625 * Fix to uni2asc() to cope with zero length Unicode strings.
15626 These are present in some PKCS#12 files.
15630 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15631 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15632 when writing a 32767 byte record.
15634 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15636 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15637 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15639 (RSA objects have a reference count access to which is protected
15640 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15641 so they are meant to be shared between threads.)
15642 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15643 "Reddie, Steven" <Steven.Reddie@ca.com>*
15645 * Fix a deadlock in CRYPTO_mem_leaks().
15649 * Use better test patterns in bntest.
15653 * rand_win.c fix for Borland C.
15657 * BN_rshift bugfix for n == 0.
15661 * Add a 'bctest' script that checks for some known 'bc' bugs
15662 so that 'make test' does not abort just because 'bc' is broken.
15666 * Store verify_result within SSL_SESSION also for client side to
15667 avoid potential security hole. (Re-used sessions on the client side
15668 always resulted in verify_result==X509_V_OK, not using the original
15669 result of the server certificate verification.)
15673 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15674 SSL3_RT_APPLICATION_DATA, return 0.
15675 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15680 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15681 releases, have been re-implemented by renaming the previous
15682 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15683 and ssl3_read_internal, respectively, and adding 'peek' parameters
15684 to them. The new ssl[23]_{read,peek} functions are calls to
15685 ssl[23]_read_internal with the 'peek' flag set appropriately.
15686 A 'peek' parameter has also been added to ssl3_read_bytes, which
15687 does the actual work for ssl3_read_internal.
15691 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15692 the method-specific "init()" handler. Also clean up ex_data after
15693 calling the method-specific "finish()" handler. Previously, this was
15694 happening the other way round.
15698 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15699 The previous value, 12, was not always sufficient for BN_mod_exp().
15703 * Make sure that shared libraries get the internal name engine with
15704 the full version number and not just 0. This should mark the
15705 shared libraries as not backward compatible. Of course, this should
15706 be changed again when we can guarantee backward binary compatibility.
15710 * Fix typo in get_cert_by_subject() in by_dir.c
15712 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15714 * Rework the system to generate shared libraries:
15716 - Make note of the expected extension for the shared libraries and
15717 if there is a need for symbolic links from for example libcrypto.so.0
15718 to libcrypto.so.0.9.7. There is extended info in Configure for
15721 - Make as few rebuilds of the shared libraries as possible.
15723 - Still avoid linking the OpenSSL programs with the shared libraries.
15725 - When installing, install the shared libraries separately from the
15730 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15732 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15733 and not in SSL_clear because the latter is also used by the
15734 accept/connect functions; previously, the settings made by
15735 SSL_set_read_ahead would be lost during the handshake.
15737 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15739 * Correct util/mkdef.pl to be selective about disabled algorithms.
15740 Previously, it would create entries for disabled algorithms no
15745 * Added several new manual pages for SSL_* function.
15749 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15751 * In ssl23_get_client_hello, generate an error message when faced
15752 with an initial SSL 3.0/TLS record that is too small to contain the
15753 first two bytes of the ClientHello message, i.e. client_version.
15754 (Note that this is a pathologic case that probably has never happened
15755 in real life.) The previous approach was to use the version number
15756 from the record header as a substitute; but our protocol choice
15757 should not depend on that one because it is not authenticated
15758 by the Finished messages.
15762 * More robust randomness gathering functions for Windows.
15764 *Jeffrey Altman <jaltman@columbia.edu>*
15766 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15767 not set then we don't setup the error code for issuer check errors
15768 to avoid possibly overwriting other errors which the callback does
15769 handle. If an application does set the flag then we assume it knows
15770 what it is doing and can handle the new informational codes
15775 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15776 a general "ANY" type, as such it should be able to decode anything
15777 including tagged types. However it didn't check the class so it would
15778 wrongly interpret tagged types in the same way as their universal
15779 counterpart and unknown types were just rejected. Changed so that the
15780 tagged and unknown types are handled in the same way as a SEQUENCE:
15781 that is the encoding is stored intact. There is also a new type
15782 "V_ASN1_OTHER" which is used when the class is not universal, in this
15783 case we have no idea what the actual type is so we just lump them all
15788 * On VMS, stdout may very well lead to a file that is written to
15789 in a record-oriented fashion. That means that every write() will
15790 write a separate record, which will be read separately by the
15791 programs trying to read from it. This can be very confusing.
15793 The solution is to put a BIO filter in the way that will buffer
15794 text until a linefeed is reached, and then write everything a
15795 line at a time, so every record written will be an actual line,
15796 not chunks of lines and not (usually doesn't happen, but I've
15797 seen it once) several lines in one record. BIO_f_linebuffer() is
15800 Currently, it's a VMS-only method, because that's where it has
15801 been tested well enough.
15805 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15806 it can return incorrect results.
15807 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15808 but it was in 0.9.6-beta[12].)
15812 * Disable the check for content being present when verifying detached
15813 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15814 include zero length content when signing messages.
15818 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15819 BIO_ctrl (for BIO pairs).
15823 * Add DSO method for VMS.
15827 * Bug fix: Montgomery multiplication could produce results with the
15832 * Add RPM specification openssl.spec and modify it to build three
15833 packages. The default package contains applications, application
15834 documentation and run-time libraries. The devel package contains
15835 include files, static libraries and function documentation. The
15836 doc package contains the contents of the doc directory. The original
15837 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15841 * Add a large number of documentation files for many SSL routines.
15843 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15845 * Add a configuration entry for Sony News 4.
15847 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15849 * Don't set the two most significant bits to one when generating a
15850 random number < q in the DSA library.
15854 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15855 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15856 the underlying transport is blocking) if a handshake took place.
15857 (The default behaviour is needed by applications such as s_client
15858 and s_server that use select() to determine when to use SSL_read;
15859 but for applications that know in advance when to expect data, it
15860 just makes things more complicated.)
15864 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15869 * Add a few more EBCDIC conditionals that make `req` and `x509`
15870 work better on such systems.
15872 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15874 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15875 Update PKCS12_parse() so it copies the friendlyName and the
15876 keyid to the certificates aux info.
15880 * Fix bug in PKCS7_verify() which caused an infinite loop
15881 if there was more than one signature.
15883 *Sven Uszpelkat <su@celocom.de>*
15885 * Major change in util/mkdef.pl to include extra information
15886 about each symbol, as well as presenting variables as well
15887 as functions. This change means that there's n more need
15888 to rebuild the .num files when some algorithms are excluded.
15892 * Allow the verify time to be set by an application,
15893 rather than always using the current time.
15897 * Phase 2 verify code reorganisation. The certificate
15898 verify code now looks up an issuer certificate by a
15899 number of criteria: subject name, authority key id
15900 and key usage. It also verifies self signed certificates
15901 by the same criteria. The main comparison function is
15902 X509_check_issued() which performs these checks.
15904 Lot of changes were necessary in order to support this
15905 without completely rewriting the lookup code.
15907 Authority and subject key identifier are now cached.
15909 The LHASH 'certs' is X509_STORE has now been replaced
15910 by a STACK_OF(X509_OBJECT). This is mainly because an
15911 LHASH can't store or retrieve multiple objects with
15912 the same hash value.
15914 As a result various functions (which were all internal
15915 use only) have changed to handle the new X509_STORE
15916 structure. This will break anything that messed round
15917 with X509_STORE internally.
15919 The functions X509_STORE_add_cert() now checks for an
15920 exact match, rather than just subject name.
15922 The X509_STORE API doesn't directly support the retrieval
15923 of multiple certificates matching a given criteria, however
15924 this can be worked round by performing a lookup first
15925 (which will fill the cache with candidate certificates)
15926 and then examining the cache for matches. This is probably
15927 the best we can do without throwing out X509_LOOKUP
15928 entirely (maybe later...).
15930 The X509_VERIFY_CTX structure has been enhanced considerably.
15932 All certificate lookup operations now go via a get_issuer()
15933 callback. Although this currently uses an X509_STORE it
15934 can be replaced by custom lookups. This is a simple way
15935 to bypass the X509_STORE hackery necessary to make this
15936 work and makes it possible to use more efficient techniques
15937 in future. A very simple version which uses a simple
15938 STACK for its trusted certificate store is also provided
15939 using X509_STORE_CTX_trusted_stack().
15941 The verify_cb() and verify() callbacks now have equivalents
15942 in the X509_STORE_CTX structure.
15944 X509_STORE_CTX also has a 'flags' field which can be used
15945 to customise the verify behaviour.
15949 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15950 excludes S/MIME capabilities.
15954 * When a certificate request is read in keep a copy of the
15955 original encoding of the signed data and use it when outputting
15956 again. Signatures then use the original encoding rather than
15957 a decoded, encoded version which may cause problems if the
15958 request is improperly encoded.
15962 * For consistency with other BIO_puts implementations, call
15963 buffer_write(b, ...) directly in buffer_puts instead of calling
15966 In BIO_puts, increment b->num_write as in BIO_write.
15968 *Peter.Sylvester@EdelWeb.fr*
15970 * Fix BN_mul_word for the case where the word is 0. (We have to use
15971 BN_zero, we may not return a BIGNUM with an array consisting of
15972 words set to zero.)
15976 * Avoid calling abort() from within the library when problems are
15977 detected, except if preprocessor symbols have been defined
15978 (such as REF_CHECK, BN_DEBUG etc.).
15982 * New openssl application 'rsautl'. This utility can be
15983 used for low-level RSA operations. DER public key
15984 BIO/fp routines also added.
15988 * New Configure entry and patches for compiling on QNX 4.
15990 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15992 * A demo state-machine implementation was sponsored by
15993 Nuron (<http://www.nuron.com/>) and is now available in
15994 demos/state_machine.
15998 * New options added to the 'dgst' utility for signature
15999 generation and verification.
16003 * Unrecognized PKCS#7 content types are now handled via a
16004 catch all ASN1_TYPE structure. This allows unsupported
16005 types to be stored as a "blob" and an application can
16006 encode and decode it manually.
16010 * Fix various signed/unsigned issues to make a_strex.c
16011 compile under VC++.
16013 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16015 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16016 length if passed a buffer. ASN1_INTEGER_to_BN failed
16017 if passed a NULL BN and its argument was negative.
16019 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16021 * Modification to PKCS#7 encoding routines to output definite
16022 length encoding. Since currently the whole structures are in
16023 memory there's not real point in using indefinite length
16024 constructed encoding. However if OpenSSL is compiled with
16025 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16029 * Added BIO_vprintf() and BIO_vsnprintf().
16033 * Added more prefixes to parse for in the strings written
16034 through a logging bio, to cover all the levels that are available
16035 through syslog. The prefixes are now:
16037 PANIC, EMERG, EMR => LOG_EMERG
16038 ALERT, ALR => LOG_ALERT
16039 CRIT, CRI => LOG_CRIT
16040 ERROR, ERR => LOG_ERR
16041 WARNING, WARN, WAR => LOG_WARNING
16042 NOTICE, NOTE, NOT => LOG_NOTICE
16043 INFO, INF => LOG_INFO
16044 DEBUG, DBG => LOG_DEBUG
16046 and as before, if none of those prefixes are present at the
16047 beginning of the string, LOG_ERR is chosen.
16049 On Win32, the `LOG_*` levels are mapped according to this:
16051 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16052 LOG_WARNING => EVENTLOG_WARNING_TYPE
16053 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16057 * Made it possible to reconfigure with just the configuration
16058 argument "reconf" or "reconfigure". The command line arguments
16059 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16060 and are retrieved from there when reconfiguring.
16066 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16068 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16072 * The obj_dat.pl script was messing up the sorting of object
16073 names. The reason was that it compared the quoted version
16074 of strings as a result "OCSP" > "OCSP Signing" because
16075 " > SPACE. Changed script to store unquoted versions of
16076 names and add quotes on output. It was also omitting some
16077 names from the lookup table if they were given a default
16078 value (that is if SN is missing it is given the same
16079 value as LN and vice versa), these are now added on the
16080 grounds that if an object has a name we should be able to
16081 look it up. Finally added warning output when duplicate
16082 short or long names are found.
16086 * Changes needed for Tandem NSK.
16088 *Scott Uroff <scott@xypro.com>*
16090 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16091 RSA_padding_check_SSLv23(), special padding was never detected
16092 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16093 version rollback attacks was not effective.
16095 In s23_clnt.c, don't use special rollback-attack detection padding
16096 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16097 client; similarly, in s23_srvr.c, don't do the rollback check if
16098 SSL 2.0 is the only protocol enabled in the server.
16102 * Make it possible to get hexdumps of unprintable data with 'openssl
16103 asn1parse'. By implication, the functions ASN1_parse_dump() and
16104 BIO_dump_indent() are added.
16108 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16109 these print out strings and name structures based on various
16110 flags including RFC2253 support and proper handling of
16111 multibyte characters. Added options to the 'x509' utility
16112 to allow the various flags to be set.
16116 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16117 Also change the functions X509_cmp_current_time() and
16118 X509_gmtime_adj() work with an ASN1_TIME structure,
16119 this will enable certificates using GeneralizedTime in validity
16120 dates to be checked.
16124 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16125 negative public key encodings) on by default,
16126 NO_NEG_PUBKEY_BUG can be set to disable it.
16130 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16131 content octets. An i2c_ASN1_OBJECT is unnecessary because
16132 the encoding can be trivially obtained from the structure.
16136 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16137 not read locks (`CRYPTO_r_[un]lock`).
16141 * A first attempt at creating official support for shared
16142 libraries through configuration. I've kept it so the
16143 default is static libraries only, and the OpenSSL programs
16144 are always statically linked for now, but there are
16145 preparations for dynamic linking in place.
16146 This has been tested on Linux and Tru64.
16150 * Randomness polling function for Win9x, as described in:
16151 Peter Gutmann, Software Generation of Practically Strong
16156 * Fix so PRNG is seeded in req if using an already existing
16161 * New options to smime application. -inform and -outform
16162 allow alternative formats for the S/MIME message including
16163 PEM and DER. The -content option allows the content to be
16164 specified separately. This should allow things like Netscape
16165 form signing output easier to verify.
16169 * Fix the ASN1 encoding of tags using the 'long form'.
16173 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16174 STRING types. These convert content octets to and from the
16175 underlying type. The actual tag and length octets are
16176 already assumed to have been read in and checked. These
16177 are needed because all other string types have virtually
16178 identical handling apart from the tag. By having versions
16179 of the ASN1 functions that just operate on content octets
16180 IMPLICIT tagging can be handled properly. It also allows
16181 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16182 and ASN1_INTEGER are identical apart from the tag.
16186 * Change the handling of OID objects as follows:
16188 - New object identifiers are inserted in objects.txt, following
16189 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16190 - objects.pl is used to process obj_mac.num and create a new
16192 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16195 This is currently kind of a hack, and the perl code in objects.pl
16196 isn't very elegant, but it works as I intended. The simplest way
16197 to check that it worked correctly is to look in obj_dat.h and
16198 check the array nid_objs and make sure the objects haven't moved
16199 around (this is important!). Additions are OK, as well as
16200 consistent name changes.
16204 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16208 * Addition of the command line parameter '-rand file' to 'openssl req'.
16209 The given file adds to whatever has already been seeded into the
16210 random pool through the RANDFILE configuration file option or
16211 environment variable, or the default random state file.
16215 * mkstack.pl now sorts each macro group into lexical order.
16216 Previously the output order depended on the order the files
16217 appeared in the directory, resulting in needless rewriting
16222 * Patches to make OpenSSL compile under Win32 again. Mostly
16223 work arounds for the VC++ problem that it treats func() as
16224 func(void). Also stripped out the parts of mkdef.pl that
16225 added extra typesafe functions: these no longer exist.
16229 * Reorganisation of the stack code. The macros are now all
16230 collected in safestack.h . Each macro is defined in terms of
16231 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16232 DEBUG_SAFESTACK is now handled in terms of function casts,
16233 this has the advantage of retaining type safety without the
16234 use of additional functions. If DEBUG_SAFESTACK is not defined
16235 then the non typesafe macros are used instead. Also modified the
16236 mkstack.pl script to handle the new form. Needs testing to see
16237 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16238 the default if no major problems. Similar behaviour for ASN1_SET_OF
16239 and PKCS12_STACK_OF.
16243 * When some versions of IIS use the 'NET' form of private key the
16244 key derivation algorithm is different. Normally MD5(password) is
16245 used as a 128 bit RC4 key. In the modified case
16246 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16247 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16248 as the old Netscape_RSA functions except they have an additional
16249 'sgckey' parameter which uses the modified algorithm. Also added
16250 an -sgckey command line option to the rsa utility. Thanks to
16251 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16252 algorithm to openssl-dev.
16256 * The evp_local.h macros were using 'c.##kname' which resulted in
16257 invalid expansion on some systems (SCO 5.0.5 for example).
16258 Corrected to 'c.kname'.
16260 *Phillip Porch <root@theporch.com>*
16262 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16263 a STACK of email addresses from a certificate or request, these look
16264 in the subject name and the subject alternative name extensions and
16265 omit any duplicate addresses.
16269 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16270 This makes DSA verification about 2 % faster.
16274 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16275 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16276 plus overhead for 1024 bit moduli).
16277 This makes exponentiations about 0.5 % faster for 1024 bit
16278 exponents (as measured by "openssl speed rsa2048").
16282 * Rename memory handling macros to avoid conflicts with other
16284 Malloc => OPENSSL_malloc
16285 Malloc_locked => OPENSSL_malloc_locked
16286 Realloc => OPENSSL_realloc
16287 Free => OPENSSL_free
16291 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16292 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16296 * CygWin32 support.
16298 *John Jarvie <jjarvie@newsguy.com>*
16300 * The type-safe stack code has been rejigged. It is now only compiled
16301 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16302 by default all type-specific stack functions are "#define"d back to
16303 standard stack functions. This results in more streamlined output
16304 but retains the type-safety checking possibilities of the original
16309 * The STACK code has been cleaned up, and certain type declarations
16310 that didn't make a lot of sense have been brought in line. This has
16311 also involved a cleanup of sorts in safestack.h to more correctly
16312 map type-safe stack functions onto their plain stack counterparts.
16313 This work has also resulted in a variety of "const"ifications of
16314 lots of the code, especially `_cmp` operations which should normally
16315 be prototyped with "const" parameters anyway.
16319 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16320 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16321 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16322 where all of 'md' is used each time the PRNG is used, but 'state'
16323 is used only indexed by a cyclic counter. As entropy may not be
16324 well distributed from the beginning, 'md' is important as a
16325 chaining variable. However, the output function chains only half
16326 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16327 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16328 in all of 'state' being rewritten, with the new values depending
16329 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16333 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16334 the handshake is continued after ssl_verify_cert_chain();
16335 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16336 can lead to 'unexplainable' connection aborts later.
16338 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16340 * Major EVP API cipher revision.
16341 Add hooks for extra EVP features. This allows various cipher
16342 parameters to be set in the EVP interface. Support added for variable
16343 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16344 setting of RC2 and RC5 parameters.
16346 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16349 Remove lots of duplicated code from the EVP library. For example *every*
16350 cipher init() function handles the 'iv' in the same way according to the
16351 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16352 for CFB and OFB modes they zero ctx->num.
16354 New functionality allows removal of S/MIME code RC2 hack.
16356 Most of the routines have the same form and so can be declared in terms
16359 By shifting this to the top level EVP_CipherInit() it can be removed from
16360 all individual ciphers. If the cipher wants to handle IVs or keys
16361 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16364 Change lots of functions like EVP_EncryptUpdate() to now return a
16365 value: although software versions of the algorithms cannot fail
16366 any installed hardware versions can.
16370 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16371 this option is set, tolerate broken clients that send the negotiated
16372 protocol version number instead of the requested protocol version
16377 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16378 i.e. non-zero for export ciphersuites, zero otherwise.
16379 Previous versions had this flag inverted, inconsistent with
16380 rsa_tmp_cb (..._TMP_RSA_CB).
16382 *Bodo Moeller; problem reported by Amit Chopra*
16384 * Add missing DSA library text string. Work around for some IIS
16385 key files with invalid SEQUENCE encoding.
16389 * Add a document (doc/standards.txt) that list all kinds of standards
16390 and so on that are implemented in OpenSSL.
16394 * Enhance c_rehash script. Old version would mishandle certificates
16395 with the same subject name hash and wouldn't handle CRLs at all.
16396 Added -fingerprint option to crl utility, to support new c_rehash
16401 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16405 * Fix for SSL server purpose checking. Server checking was
16406 rejecting certificates which had extended key usage present
16407 but no ssl client purpose.
16409 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16411 * Make PKCS#12 code work with no password. The PKCS#12 spec
16412 is a little unclear about how a blank password is handled.
16413 Since the password in encoded as a BMPString with terminating
16414 double NULL a zero length password would end up as just the
16415 double NULL. However no password at all is different and is
16416 handled differently in the PKCS#12 key generation code. NS
16417 treats a blank password as zero length. MSIE treats it as no
16418 password on export: but it will try both on import. We now do
16419 the same: PKCS12_parse() tries zero length and no password if
16420 the password is set to "" or NULL (NULL is now a valid password:
16421 it wasn't before) as does the pkcs12 application.
16425 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16426 perror when PEM_read_bio_X509_REQ fails, the error message must
16427 be obtained from the error queue.
16431 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16432 it in ERR_remove_state if appropriate, and change ERR_get_state
16433 accordingly to avoid race conditions (this is necessary because
16434 thread_hash is no longer constant once set).
16438 * Bugfix for linux-elf makefile.one.
16442 * RSA_get_default_method() will now cause a default
16443 RSA_METHOD to be chosen if one doesn't exist already.
16444 Previously this was only set during a call to RSA_new()
16445 or RSA_new_method(NULL) meaning it was possible for
16446 RSA_get_default_method() to return NULL.
16450 * Added native name translation to the existing DSO code
16451 that will convert (if the flag to do so is set) filenames
16452 that are sufficiently small and have no path information
16453 into a canonical native form. Eg. "blah" converted to
16454 "libblah.so" or "blah.dll" etc.
16458 * New function ERR_error_string_n(e, buf, len) which is like
16459 ERR_error_string(e, buf), but writes at most 'len' bytes
16460 including the 0 terminator. For ERR_error_string_n, 'buf'
16463 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16465 * CONF library reworked to become more general. A new CONF
16466 configuration file reader "class" is implemented as well as a
16467 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16468 old `CONF_*` functions are still there, but are reimplemented to
16469 work in terms of the new functions. Also, a set of functions
16470 to handle the internal storage of the configuration data is
16471 provided to make it easier to write new configuration file
16472 reader "classes" (I can definitely see something reading a
16473 configuration file in XML format, for example), called `_CONF_*`,
16474 or "the configuration storage API"...
16476 The new configuration file reading functions are:
16478 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16479 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16481 NCONF_default, NCONF_WIN32
16483 NCONF_dump_fp, NCONF_dump_bio
16485 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16486 NCONF_new creates a new CONF object. This works in the same way
16487 as other interfaces in OpenSSL, like the BIO interface.
16488 `NCONF_dump_*` dump the internal storage of the configuration file,
16489 which is useful for debugging. All other functions take the same
16490 arguments as the old `CONF_*` functions with the exception of the
16491 first that must be a `CONF *` instead of a `LHASH *`.
16493 To make it easier to use the new classes with the old `CONF_*` functions,
16494 the function CONF_set_default_method is provided.
16498 * Add '-tls1' option to 'openssl ciphers', which was already
16499 mentioned in the documentation but had not been implemented.
16500 (This option is not yet really useful because even the additional
16501 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16505 * Initial DSO code added into libcrypto for letting OpenSSL (and
16506 OpenSSL-based applications) load shared libraries and bind to
16507 them in a portable way.
16509 *Geoff Thorpe, with contributions from Richard Levitte*
16511 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16513 * Make sure _lrotl and _lrotr are only used with MSVC.
16515 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16516 (the default implementation of RAND_status).
16518 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16519 to '-clrext' (= clear extensions), as intended and documented.
16520 *Bodo Moeller; inconsistency pointed out by Michael Attili
16521 <attili@amaxo.com>*
16523 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16524 was larger than the MD block size.
16526 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16528 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16529 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16530 using the passed key: if the passed key was a private key the result
16531 of X509_print(), for example, would be to print out all the private key
16536 * des_quad_cksum() byte order bug fix.
16537 *Ulf Möller, using the problem description in krb4-0.9.7, where
16538 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16540 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16543 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16545 * For easily testing in shell scripts whether some command
16546 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16547 returns with exit code 0 iff no command of the given name is available.
16548 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16549 the output goes to stdout and nothing is printed to stderr.
16550 Additional arguments are always ignored.
16552 Since for each cipher there is a command of the same name,
16553 the 'no-cipher' compilation switches can be tested this way.
16555 ('openssl no-XXX' is not able to detect pseudo-commands such
16556 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16560 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16564 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16565 is set; it will be thrown away anyway because each handshake creates
16567 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16568 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16569 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16570 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16574 * New s_client option -ign_eof: EOF at stdin is ignored, and
16575 'Q' and 'R' lose their special meanings (quit/renegotiate).
16576 This is part of what -quiet does; unlike -quiet, -ign_eof
16577 does not suppress any output.
16581 * Add compatibility options to the purpose and trust code. The
16582 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16583 accepts a certificate or CA, this was the previous behaviour,
16584 with all the associated security issues.
16586 X509_TRUST_COMPAT is the old trust behaviour: only and
16587 automatically trust self signed roots in certificate store. A
16588 new trust setting X509_TRUST_DEFAULT is used to specify that
16589 a purpose has no associated trust setting and it should instead
16590 use the value in the default purpose.
16594 * Fix the PKCS#8 DSA private key code so it decodes keys again
16595 and fix a memory leak.
16599 * In util/mkerr.pl (which implements 'make errors'), preserve
16600 reason strings from the previous version of the .c file, as
16601 the default to have only downcase letters (and digits) in
16602 automatically generated reasons codes is not always appropriate.
16606 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16607 using strerror. Previously, ERR_reason_error_string() returned
16608 library names as reason strings for SYSerr; but SYSerr is a special
16609 case where small numbers are errno values, not library numbers.
16613 * Add '-dsaparam' option to 'openssl dhparam' application. This
16614 converts DSA parameters into DH parameters. (When creating parameters,
16615 DSA_generate_parameters is used.)
16619 * Include 'length' (recommended exponent length) in C code generated
16620 by 'openssl dhparam -C'.
16624 * The second argument to set_label in perlasm was already being used
16625 so couldn't be used as a "file scope" flag. Moved to third argument
16630 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16631 instead of RAND_bytes for encryption IVs and salts.
16635 * Include RAND_status() into RAND_METHOD instead of implementing
16636 it only for md_rand.c Otherwise replacing the PRNG by calling
16637 RAND_set_rand_method would be impossible.
16641 * Don't let DSA_generate_key() enter an infinite loop if the random
16642 number generation fails.
16646 * New 'rand' application for creating pseudo-random output.
16650 * Added configuration support for Linux/IA64
16652 *Rolf Haberrecker <rolf@suse.de>*
16654 * Assembler module support for Mingw32.
16658 * Shared library support for HPUX (in shlib/).
16660 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16662 * Shared library support for Solaris gcc.
16664 *Lutz Behnke <behnke@trustcenter.de>*
16666 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16668 * PKCS7_encrypt() was adding text MIME headers twice because they
16669 were added manually and by SMIME_crlf_copy().
16673 * In bntest.c don't call BN_rand with zero bits argument.
16675 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16677 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16678 case was implemented. This caused BN_div_recp() to fail occasionally.
16682 * Add an optional second argument to the set_label() in the perl
16683 assembly language builder. If this argument exists and is set
16684 to 1 it signals that the assembler should use a symbol whose
16685 scope is the entire file, not just the current function. This
16686 is needed with MASM which uses the format label:: for this scope.
16688 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16690 * Change the ASN1 types so they are typedefs by default. Before
16691 almost all types were #define'd to ASN1_STRING which was causing
16692 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16697 * Change names of new functions to the new get1/get0 naming
16698 convention: After 'get1', the caller owns a reference count
16699 and has to call `..._free`; 'get0' returns a pointer to some
16700 data structure without incrementing reference counters.
16701 (Some of the existing 'get' functions increment a reference
16702 counter, some don't.)
16703 Similarly, 'set1' and 'add1' functions increase reference
16704 counters or duplicate objects.
16708 * Allow for the possibility of temp RSA key generation failure:
16709 the code used to assume it always worked and crashed on failure.
16713 * Fix potential buffer overrun problem in BIO_printf().
16714 *Ulf Möller, using public domain code by Patrick Powell; problem
16715 pointed out by David Sacerdote <das33@cornell.edu>*
16717 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16718 RAND_egd() and RAND_status(). In the command line application,
16719 the EGD socket can be specified like a seed file using RANDFILE
16724 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16725 Some CAs (e.g. Verisign) distribute certificates in this form.
16729 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16730 list to exclude them. This means that no special compilation option
16731 is needed to use anonymous DH: it just needs to be included in the
16736 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16737 EVP_MD_type. The old functionality is available in a new macro called
16738 EVP_MD_md(). Change code that uses it and update docs.
16742 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16743 where the `void *` argument is replaced by a function pointer argument.
16744 Previously `void *` was abused to point to functions, which works on
16745 many platforms, but is not correct. As these functions are usually
16746 called by macros defined in OpenSSL header files, most source code
16747 should work without changes.
16751 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16752 sections with information on -D... compiler switches used for
16753 compiling the library so that applications can see them. To enable
16754 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16755 must be defined. E.g.,
16756 #define OPENSSL_ALGORITHM_DEFINES
16757 #include <openssl/opensslconf.h>
16758 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16760 *Richard Levitte, Ulf and Bodo Möller*
16762 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16767 * Change the 'other' type in certificate aux info to a STACK_OF
16768 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16769 the required ASN1 format: arbitrary types determined by an OID.
16773 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16774 argument to 'req'. This is not because the function is newer or
16775 better than others it just uses the work 'NEW' in the certificate
16776 request header lines. Some software needs this.
16780 * Reorganise password command line arguments: now passwords can be
16781 obtained from various sources. Delete the PEM_cb function and make
16782 it the default behaviour: i.e. if the callback is NULL and the
16783 usrdata argument is not NULL interpret it as a null terminated pass
16784 phrase. If usrdata and the callback are NULL then the pass phrase
16785 is prompted for as usual.
16789 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16790 the support is automatically enabled. The resulting binaries will
16791 autodetect the card and use it if present.
16793 *Ben Laurie and Compaq Inc.*
16795 * Work around for Netscape hang bug. This sends certificate request
16796 and server done in one record. Since this is perfectly legal in the
16797 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16798 the bugs/SSLv3 entry for more info.
16802 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16806 * Add -rand argument to smime and pkcs12 applications and read/write
16811 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16815 * Add command line password options to the remaining applications.
16819 * Bug fix for BN_div_recp() for numerators with an even number of
16824 * More tests in bntest.c, and changed test_bn output.
16828 * ./config recognizes MacOS X now.
16832 * Bug fix for BN_div() when the first words of num and divisor are
16833 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16837 * Add support for various broken PKCS#8 formats, and command line
16838 options to produce them.
16842 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16843 get temporary BIGNUMs from a BN_CTX.
16847 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16852 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16853 include a #define from the old name to the new. The original intent
16854 was that statically linked binaries could for example just call
16855 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16856 link with digests. This never worked because SSLeay_add_all_digests()
16857 and SSLeay_add_all_ciphers() were in the same source file so calling
16858 one would link with the other. They are now in separate source files.
16862 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16866 * Use a less unusual form of the Miller-Rabin primality test (it used
16867 a binary algorithm for exponentiation integrated into the Miller-Rabin
16868 loop, our standard modexp algorithms are faster).
16872 * Support for the EBCDIC character set completed.
16874 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16876 * Source code cleanups: use const where appropriate, eliminate casts,
16877 use `void *` instead of `char *` in lhash.
16881 * Bugfix: ssl3_send_server_key_exchange was not restartable
16882 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16883 this the server could overwrite ephemeral keys that the client
16888 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16889 using 50 iterations of the Rabin-Miller test.
16891 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16892 iterations of the Rabin-Miller test as required by the appendix
16893 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16894 As BN_is_prime_fasttest includes trial division, DSA parameter
16895 generation becomes much faster.
16897 This implies a change for the callback functions in DSA_is_prime
16898 and DSA_generate_parameters: The callback function is called once
16899 for each positive witness in the Rabin-Miller test, not just
16900 occasionally in the inner loop; and the parameters to the
16901 callback function now provide an iteration count for the outer
16902 loop rather than for the current invocation of the inner loop.
16903 DSA_generate_parameters additionally can call the callback
16904 function with an 'iteration count' of -1, meaning that a
16905 candidate has passed the trial division test (when q is generated
16906 from an application-provided seed, trial division is skipped).
16910 * New function BN_is_prime_fasttest that optionally does trial
16911 division before starting the Rabin-Miller test and has
16912 an additional BN_CTX * argument (whereas BN_is_prime always
16913 has to allocate at least one BN_CTX).
16914 'callback(1, -1, cb_arg)' is called when a number has passed the
16915 trial division stage.
16919 * Fix for bug in CRL encoding. The validity dates weren't being handled
16924 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16928 * New function BN_pseudo_rand().
16932 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16933 bignum version of BN_from_montgomery() with the working code from
16934 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16939 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16940 made it impossible to use the same SSL_SESSION data structure in
16941 SSL2 clients in multiple threads.
16945 * The return value of RAND_load_file() no longer counts bytes obtained
16946 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16947 to seed the PRNG (previously an explicit byte count was required).
16949 *Ulf Möller, Bodo Möller*
16951 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16952 used `char *` instead of `void *` and had casts all over the place.
16956 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16960 * Retain source code compatibility for BN_prime_checks macro:
16961 BN_is_prime(..., BN_prime_checks, ...) now uses
16962 BN_prime_checks_for_size to determine the appropriate number of
16963 Rabin-Miller iterations.
16967 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16968 DH_CHECK_P_NOT_SAFE_PRIME.
16969 (Check if this is true? OpenPGP calls them "strong".)
16973 * Merge the functionality of "dh" and "gendh" programs into a new program
16974 "dhparam". The old programs are retained for now but will handle DH keys
16975 (instead of parameters) in future.
16979 * Make the ciphers, s_server and s_client programs check the return values
16980 when a new cipher list is set.
16984 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16985 ciphers. Before when the 56bit ciphers were enabled the sorting was
16988 The syntax for the cipher sorting has been extended to support sorting by
16989 cipher-strength (using the strength_bits hard coded in the tables).
16990 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16992 Fix a bug in the cipher-command parser: when supplying a cipher command
16993 string with an "undefined" symbol (neither command nor alphanumeric
16994 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16995 an error is flagged.
16997 Due to the strength-sorting extension, the code of the
16998 ssl_create_cipher_list() function was completely rearranged. I hope that
16999 the readability was also increased :-)
17001 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17003 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17004 for the first serial number and places 2 in the serial number file. This
17005 avoids problems when the root CA is created with serial number zero and
17006 the first user certificate has the same issuer name and serial number
17011 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17012 the new code. Add documentation for this stuff.
17016 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17017 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17018 structures and behave in an analogous way to the X509v3 functions:
17019 they shouldn't be called directly but wrapper functions should be used
17022 So we also now have some wrapper functions that call the X509at functions
17023 when passed certificate requests. (TO DO: similar things can be done with
17024 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17025 things. Some of these need some d2i or i2d and print functionality
17026 because they handle more complex structures.)
17030 * Add missing #ifndefs that caused missing symbols when building libssl
17031 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17032 NO_RSA in `ssl/s2*.c`.
17034 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17036 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17037 has a return value which indicates the quality of the random data
17038 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17039 error queue. New function RAND_pseudo_bytes() generates output that is
17040 guaranteed to be unique but not unpredictable. RAND_add is like
17041 RAND_seed, but takes an extra argument for an entropy estimate
17042 (RAND_seed always assumes full entropy).
17046 * Do more iterations of Rabin-Miller probable prime test (specifically,
17047 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17048 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17049 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17050 false-positive rate of at most 2^-80 for random input.
17054 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17058 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17059 in the 0.9.5 release), this returns the chain
17060 from an X509_CTX structure with a dup of the stack and all
17061 the X509 reference counts upped: so the stack will exist
17062 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17065 Also make SSL_SESSION_print() print out the verify return
17070 * Add manpage for the pkcs12 command. Also change the default
17071 behaviour so MAC iteration counts are used unless the new
17072 -nomaciter option is used. This improves file security and
17073 only older versions of MSIE (4.0 for example) need it.
17077 * Honor the no-xxx Configure options when creating .DEF files.
17081 * Add PKCS#10 attributes to field table: challengePassword,
17082 unstructuredName and unstructuredAddress. These are taken from
17083 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17084 international characters are used.
17086 More changes to X509_ATTRIBUTE code: allow the setting of types
17087 based on strings. Remove the 'loc' parameter when adding
17088 attributes because these will be a SET OF encoding which is sorted
17093 * Initial changes to the 'req' utility to allow request generation
17094 automation. This will allow an application to just generate a template
17095 file containing all the field values and have req construct the
17098 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17099 used all over the place including certificate requests and PKCS#7
17100 structures. They are currently handled manually where necessary with
17101 some primitive wrappers for PKCS#7. The new functions behave in a
17102 manner analogous to the X509 extension functions: they allow
17103 attributes to be looked up by NID and added.
17105 Later something similar to the X509V3 code would be desirable to
17106 automatically handle the encoding, decoding and printing of the
17107 more complex types. The string types like challengePassword can
17108 be handled by the string table functions.
17110 Also modified the multi byte string table handling. Now there is
17111 a 'global mask' which masks out certain types. The table itself
17112 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17113 is useful when for example there is only one permissible type
17114 (as in countryName) and using the mask might result in no valid
17119 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17120 SSL_get_peer_finished to allow applications to obtain the latest
17121 Finished messages sent to the peer or expected from the peer,
17122 respectively. (SSL_get_peer_finished is usually the Finished message
17123 actually received from the peer, otherwise the protocol will be aborted.)
17125 As the Finished message are message digests of the complete handshake
17126 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17127 be used for external authentication procedures when the authentication
17128 provided by SSL/TLS is not desired or is not enough.
17132 * Enhanced support for Alpha Linux is added. Now ./config checks if
17133 the host supports BWX extension and if Compaq C is present on the
17134 $PATH. Just exploiting of the BWX extension results in 20-30%
17135 performance kick for some algorithms, e.g. DES and RC4 to mention
17136 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17141 * Add support for MS "fast SGC". This is arguably a violation of the
17142 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17143 weak crypto and after checking the certificate is SGC a second one
17144 with strong crypto. MS SGC stops the first handshake after receiving
17145 the server certificate message and sends a second client hello. Since
17146 a server will typically do all the time consuming operations before
17147 expecting any further messages from the client (server key exchange
17148 is the most expensive) there is little difference between the two.
17150 To get OpenSSL to support MS SGC we have to permit a second client
17151 hello message after we have sent server done. In addition we have to
17152 reset the MAC if we do get this second client hello.
17156 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17157 if a DER encoded private key is RSA or DSA traditional format. Changed
17158 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17159 format DER encoded private key. Newer code should use PKCS#8 format which
17160 has the key type encoded in the ASN1 structure. Added DER private key
17161 support to pkcs8 application.
17165 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17166 ciphersuites has been selected (as required by the SSL 3/TLS 1
17167 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17168 is set, we interpret this as a request to violate the specification
17169 (the worst that can happen is a handshake failure, and 'correct'
17170 behaviour would result in a handshake failure anyway).
17174 * In SSL_CTX_add_session, take into account that there might be multiple
17175 SSL_SESSION structures with the same session ID (e.g. when two threads
17176 concurrently obtain them from an external cache).
17177 The internal cache can handle only one SSL_SESSION with a given ID,
17178 so if there's a conflict, we now throw out the old one to achieve
17183 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17184 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17185 some routines that use cipher OIDs: some ciphers do not have OIDs
17186 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17191 * Simplify the trust setting structure and code. Now we just have
17192 two sequences of OIDs for trusted and rejected settings. These will
17193 typically have values the same as the extended key usage extension
17194 and any application specific purposes.
17196 The trust checking code now has a default behaviour: it will just
17197 check for an object with the same NID as the passed id. Functions can
17198 be provided to override either the default behaviour or the behaviour
17199 for a given id. SSL client, server and email already have functions
17200 in place for compatibility: they check the NID and also return "trusted"
17201 if the certificate is self signed.
17205 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17206 traditional format into an EVP_PKEY structure.
17210 * Add a password callback function PEM_cb() which either prompts for
17211 a password if usr_data is NULL or otherwise assumes it is a null
17212 terminated password. Allow passwords to be passed on command line
17213 environment or config files in a few more utilities.
17217 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17218 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17219 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17220 Update documentation.
17224 * Support for ASN1 "NULL" type. This could be handled before by using
17225 ASN1_TYPE but there wasn't any function that would try to read a NULL
17226 and produce an error if it couldn't. For compatibility we also have
17227 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17228 don't allocate anything because they don't need to.
17232 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17235 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17237 * Rebuild of the memory allocation routines used by OpenSSL code and
17238 possibly others as well. The purpose is to make an interface that
17239 provide hooks so anyone can build a separate set of allocation and
17240 deallocation routines to be used by OpenSSL, for example memory
17241 pool implementations, or something else, which was previously hard
17242 since Malloc(), Realloc() and Free() were defined as macros having
17243 the values malloc, realloc and free, respectively (except for Win32
17244 compilations). The same is provided for memory debugging code.
17245 OpenSSL already comes with functionality to find memory leaks, but
17246 this gives people a chance to debug other memory problems.
17248 With these changes, a new set of functions and macros have appeared:
17250 CRYPTO_set_mem_debug_functions() [F]
17251 CRYPTO_get_mem_debug_functions() [F]
17252 CRYPTO_dbg_set_options() [F]
17253 CRYPTO_dbg_get_options() [F]
17254 CRYPTO_malloc_debug_init() [M]
17256 The memory debug functions are NULL by default, unless the library
17257 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17258 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17259 gives the standard debugging functions that come with OpenSSL) or
17260 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17261 provided by the library user) must be used. When the standard
17262 debugging functions are used, CRYPTO_dbg_set_options can be used to
17263 request additional information:
17264 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17265 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17267 Also, things like CRYPTO_set_mem_functions will always give the
17268 expected result (the new set of functions is used for allocation
17269 and deallocation) at all times, regardless of platform and compiler
17272 To finish it up, some functions that were never use in any other
17273 way than through macros have a new API and new semantic:
17275 CRYPTO_dbg_malloc()
17276 CRYPTO_dbg_realloc()
17279 All macros of value have retained their old syntax.
17281 *Richard Levitte and Bodo Moeller*
17283 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17284 ordering of SMIMECapabilities wasn't in "strength order" and there
17285 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17290 * Some ASN1 types with illegal zero length encoding (INTEGER,
17291 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17293 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17295 * Merge in my S/MIME library for OpenSSL. This provides a simple
17296 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17297 functionality to handle multipart/signed properly) and a utility
17298 called 'smime' to call all this stuff. This is based on code I
17299 originally wrote for Celo who have kindly allowed it to be
17300 included in OpenSSL.
17304 * Add variants des_set_key_checked and des_set_key_unchecked of
17305 des_set_key (aka des_key_sched). Global variable des_check_key
17306 decides which of these is called by des_set_key; this way
17307 des_check_key behaves as it always did, but applications and
17308 the library itself, which was buggy for des_check_key == 1,
17309 have a cleaner way to pick the version they need.
17313 * New function PKCS12_newpass() which changes the password of a
17318 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17319 dynamic mix. In both cases the ids can be used as an index into the
17320 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17321 functions so they accept a list of the field values and the
17322 application doesn't need to directly manipulate the X509_TRUST
17327 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17332 * Modify the way the V3 extension code looks up extensions. This now
17333 works in a similar way to the object code: we have some "standard"
17334 extensions in a static table which is searched with OBJ_bsearch()
17335 and the application can add dynamic ones if needed. The file
17336 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17337 updated whenever a new extension is added to the core code and kept
17338 in ext_nid order. There is a simple program 'tabtest.c' which checks
17339 this. New extensions are not added too often so this file can readily
17340 be maintained manually.
17342 There are two big advantages in doing things this way. The extensions
17343 can be looked up immediately and no longer need to be "added" using
17344 X509V3_add_standard_extensions(): this function now does nothing.
17345 Side note: I get *lots* of email saying the extension code doesn't
17346 work because people forget to call this function.
17347 Also no dynamic allocation is done unless new extensions are added:
17348 so if we don't add custom extensions there is no need to call
17349 X509V3_EXT_cleanup().
17353 * Modify enc utility's salting as follows: make salting the default. Add a
17354 magic header, so unsalted files fail gracefully instead of just decrypting
17355 to garbage. This is because not salting is a big security hole, so people
17356 should be discouraged from doing it.
17360 * Fixes and enhancements to the 'x509' utility. It allowed a message
17361 digest to be passed on the command line but it only used this
17362 parameter when signing a certificate. Modified so all relevant
17363 operations are affected by the digest parameter including the
17364 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17365 DSA key was used because it didn't fix the digest.
17369 * Initial certificate chain verify code. Currently tests the untrusted
17370 certificates for consistency with the verify purpose (which is set
17371 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17373 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17374 this is because it will reject chains with invalid extensions whereas
17375 every previous version of OpenSSL and SSLeay made no checks at all.
17377 Trust code: checks the root CA for the relevant trust settings. Trust
17378 settings have an initial value consistent with the verify purpose: e.g.
17379 if the verify purpose is for SSL client use it expects the CA to be
17380 trusted for SSL client use. However the default value can be changed to
17381 permit custom trust settings: one example of this would be to only trust
17382 certificates from a specific "secure" set of CAs.
17384 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17385 which should be used for version portability: especially since the
17386 verify structure is likely to change more often now.
17388 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17389 to set them. If not set then assume SSL clients will verify SSL servers
17392 Two new options to the verify program: -untrusted allows a set of
17393 untrusted certificates to be passed in and -purpose which sets the
17394 intended purpose of the certificate. If a purpose is set then the
17395 new chain verify code is used to check extension consistency.
17399 * Support for the authority information access extension.
17403 * Modify RSA and DSA PEM read routines to transparently handle
17404 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17405 public keys in a format compatible with certificate
17406 SubjectPublicKeyInfo structures. Unfortunately there were already
17407 functions called *_PublicKey_* which used various odd formats so
17408 these are retained for compatibility: however the DSA variants were
17409 never in a public release so they have been deleted. Changed dsa/rsa
17410 utilities to handle the new format: note no releases ever handled public
17411 keys so we should be OK.
17413 The primary motivation for this change is to avoid the same fiasco
17414 that dogs private keys: there are several incompatible private key
17415 formats some of which are standard and some OpenSSL specific and
17416 require various evil hacks to allow partial transparent handling and
17417 even then it doesn't work with DER formats. Given the option anything
17418 other than PKCS#8 should be dumped: but the other formats have to
17419 stay in the name of compatibility.
17421 With public keys and the benefit of hindsight one standard format
17422 is used which works with EVP_PKEY, RSA or DSA structures: though
17423 it clearly returns an error if you try to read the wrong kind of key.
17425 Added a -pubkey option to the 'x509' utility to output the public key.
17426 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17427 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17428 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17429 that do the same as the `EVP_PKEY_assign_*()` except they up the
17430 reference count of the added key (they don't "swallow" the
17435 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17436 CRLs would fail if the file contained no certificates or no CRLs:
17437 added a new function to read in both types and return the number
17438 read: this means that if none are read it will be an error. The
17439 DER versions of the certificate and CRL reader would always fail
17440 because it isn't possible to mix certificates and CRLs in DER format
17441 without choking one or the other routine. Changed this to just read
17442 a certificate: this is the best we can do. Also modified the code
17443 in `apps/verify.c` to take notice of return codes: it was previously
17444 attempting to read in certificates from NULL pointers and ignoring
17445 any errors: this is one reason why the cert and CRL reader seemed
17446 to work. It doesn't check return codes from the default certificate
17447 routines: these may well fail if the certificates aren't installed.
17451 * Code to support otherName option in GeneralName.
17455 * First update to verify code. Change the verify utility
17456 so it warns if it is passed a self signed certificate:
17457 for consistency with the normal behaviour. X509_verify
17458 has been modified to it will now verify a self signed
17459 certificate if *exactly* the same certificate appears
17460 in the store: it was previously impossible to trust a
17461 single self signed certificate. This means that:
17462 openssl verify ss.pem
17463 now gives a warning about a self signed certificate but
17464 openssl verify -CAfile ss.pem ss.pem
17469 * For servers, store verify_result in SSL_SESSION data structure
17470 (and add it to external session representation).
17471 This is needed when client certificate verifications fails,
17472 but an application-provided verification callback (set by
17473 SSL_CTX_set_cert_verify_callback) allows accepting the session
17474 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17475 but returns 1): When the session is reused, we have to set
17476 ssl->verify_result to the appropriate error code to avoid
17479 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17481 * Fix a bug in the new PKCS#7 code: it didn't consider the
17482 case in PKCS7_dataInit() where the signed PKCS7 structure
17483 didn't contain any existing data because it was being created.
17485 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17487 * Add a salt to the key derivation routines in enc.c. This
17488 forms the first 8 bytes of the encrypted file. Also add a
17489 -S option to allow a salt to be input on the command line.
17493 * New function X509_cmp(). Oddly enough there wasn't a function
17494 to compare two certificates. We do this by working out the SHA1
17495 hash and comparing that. X509_cmp() will be needed by the trust
17500 * SSL_get1_session() is like SSL_get_session(), but increments
17501 the reference count in the SSL_SESSION returned.
17503 *Geoff Thorpe <geoff@eu.c2.net>*
17505 * Fix for 'req': it was adding a null to request attributes.
17506 Also change the X509_LOOKUP and X509_INFO code to handle
17507 certificate auxiliary information.
17511 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17516 * Add the possibility to add extra information to the memory leak
17517 detecting output, to form tracebacks, showing from where each
17518 allocation was originated: CRYPTO_push_info("constant string") adds
17519 the string plus current file name and line number to a per-thread
17520 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17521 is like calling CYRPTO_pop_info() until the stack is empty.
17522 Also updated memory leak detection code to be multi-thread-safe.
17526 * Add options -text and -noout to pkcs7 utility and delete the
17527 encryption options which never did anything. Update docs.
17531 * Add options to some of the utilities to allow the pass phrase
17532 to be included on either the command line (not recommended on
17533 OSes like Unix) or read from the environment. Update the
17534 manpages and fix a few bugs.
17538 * Add a few manpages for some of the openssl commands.
17542 * Fix the -revoke option in ca. It was freeing up memory twice,
17543 leaking and not finding already revoked certificates.
17547 * Extensive changes to support certificate auxiliary information.
17548 This involves the use of X509_CERT_AUX structure and X509_AUX
17549 functions. An X509_AUX function such as PEM_read_X509_AUX()
17550 can still read in a certificate file in the usual way but it
17551 will also read in any additional "auxiliary information". By
17552 doing things this way a fair degree of compatibility can be
17553 retained: existing certificates can have this information added
17554 using the new 'x509' options.
17556 Current auxiliary information includes an "alias" and some trust
17557 settings. The trust settings will ultimately be used in enhanced
17558 certificate chain verification routines: currently a certificate
17559 can only be trusted if it is self signed and then it is trusted
17564 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17565 The problem was that one of the replacement routines had not been working
17566 since SSLeay releases. For now the offending routine has been replaced
17567 with non-optimised assembler. Even so, this now gives around 95%
17568 performance improvement for 1024 bit RSA signs.
17572 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17573 handling. Most clients have the effective key size in bits equal to
17574 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17575 A few however don't do this and instead use the size of the decrypted key
17576 to determine the RC2 key length and the AlgorithmIdentifier to determine
17577 the effective key length. In this case the effective key length can still
17578 be 40 bits but the key length can be 168 bits for example. This is fixed
17579 by manually forcing an RC2 key into the EVP_PKEY structure because the
17580 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17581 the key length and effective key length are equal.
17585 * Add a bunch of functions that should simplify the creation of
17586 X509_NAME structures. Now you should be able to do:
17587 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17588 and have it automatically work out the correct field type and fill in
17589 the structures. The more adventurous can try:
17590 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17591 and it will (hopefully) work out the correct multibyte encoding.
17595 * Change the 'req' utility to use the new field handling and multibyte
17596 copy routines. Before the DN field creation was handled in an ad hoc
17597 way in req, ca, and x509 which was rather broken and didn't support
17598 BMPStrings or UTF8Strings. Since some software doesn't implement
17599 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17600 using the dirstring_type option. See the new comment in the default
17601 openssl.cnf for more info.
17605 * Make crypto/rand/md_rand.c more robust:
17606 - Assure unique random numbers after fork().
17607 - Make sure that concurrent threads access the global counter and
17608 md serializably so that we never lose entropy in them
17609 or use exactly the same state in multiple threads.
17610 Access to the large state is not always serializable because
17611 the additional locking could be a performance killer, and
17612 md should be large enough anyway.
17616 * New file `apps/app_rand.c` with commonly needed functionality
17617 for handling the random seed file.
17619 Use the random seed file in some applications that previously did not:
17621 dsaparam -genkey (which also ignored its '-rand' option),
17624 x509 (when signing).
17625 Except on systems with /dev/urandom, it is crucial to have a random
17626 seed file at least for key creation, DSA signing, and for DH exchanges;
17627 for RSA signatures we could do without one.
17629 gendh and gendsa (unlike genrsa) used to read only the first byte
17630 of each file listed in the '-rand' option. The function as previously
17631 found in genrsa is now in app_rand.c and is used by all programs
17632 that support '-rand'.
17636 * In RAND_write_file, use mode 0600 for creating files;
17637 don't just chmod when it may be too late.
17641 * Report an error from X509_STORE_load_locations
17642 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17646 * New function ASN1_mbstring_copy() this copies a string in either
17647 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17648 into an ASN1_STRING type. A mask of permissible types is passed
17649 and it chooses the "minimal" type to use or an error if not type
17654 * Add function equivalents to the various macros in asn1.h. The old
17655 macros are retained with an `M_` prefix. Code inside the library can
17656 use the `M_` macros. External code (including the openssl utility)
17657 should *NOT* in order to be "shared library friendly".
17661 * Add various functions that can check a certificate's extensions
17662 to see if it usable for various purposes such as SSL client,
17663 server or S/MIME and CAs of these types. This is currently
17664 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17665 verification. Also added a -purpose flag to x509 utility to
17666 print out all the purposes.
17670 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17675 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17676 for, obtain and decode and extension and obtain its critical flag.
17677 This allows all the necessary extension code to be handled in a
17678 single function call.
17682 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17683 platforms. See crypto/rc4/rc4_enc.c for further details.
17687 * New -noout option to asn1parse. This causes no output to be produced
17688 its main use is when combined with -strparse and -out to extract data
17689 from a file (which may not be in ASN.1 format).
17693 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17694 when producing the local key id.
17696 *Richard Levitte <levitte@stacken.kth.se>*
17698 * New option -dhparam in s_server. This allows a DH parameter file to be
17699 stated explicitly. If it is not stated then it tries the first server
17700 certificate file. The previous behaviour hard coded the filename
17705 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17706 a public key to be input or output. For example:
17707 openssl rsa -in key.pem -pubout -out pubkey.pem
17708 Also added necessary DSA public key functions to handle this.
17712 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17713 in the message. This was handled by allowing
17714 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17716 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17718 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17719 to the end of the strings whereas this didn't. This would cause problems
17720 if strings read with d2i_ASN1_bytes() were later modified.
17722 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17724 * Fix for base64 decode bug. When a base64 bio reads only one line of
17725 data and it contains EOF it will end up returning an error. This is
17726 caused by input 46 bytes long. The cause is due to the way base64
17727 BIOs find the start of base64 encoded data. They do this by trying a
17728 trial decode on each line until they find one that works. When they
17729 do a flag is set and it starts again knowing it can pass all the
17730 data directly through the decoder. Unfortunately it doesn't reset
17731 the context it uses. This means that if EOF is reached an attempt
17732 is made to pass two EOFs through the context and this causes the
17733 resulting error. This can also cause other problems as well. As is
17734 usual with these problems it takes *ages* to find and the fix is
17735 trivial: move one line.
17737 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17739 * Ugly workaround to get s_client and s_server working under Windows. The
17740 old code wouldn't work because it needed to select() on sockets and the
17741 tty (for keypresses and to see if data could be written). Win32 only
17742 supports select() on sockets so we select() with a 1s timeout on the
17743 sockets and then see if any characters are waiting to be read, if none
17744 are present then we retry, we also assume we can always write data to
17745 the tty. This isn't nice because the code then blocks until we've
17746 received a complete line of data and it is effectively polling the
17747 keyboard at 1s intervals: however it's quite a bit better than not
17748 working at all :-) A dedicated Windows application might handle this
17749 with an event loop for example.
17753 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17754 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17755 will be called when RSA_sign() and RSA_verify() are used. This is useful
17756 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17757 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17758 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17759 This necessitated the support of an extra signature type NID_md5_sha1
17760 for SSL signatures and modifications to the SSL library to use it instead
17761 of calling RSA_public_decrypt() and RSA_private_encrypt().
17765 * Add new -verify -CAfile and -CApath options to the crl program, these
17766 will lookup a CRL issuers certificate and verify the signature in a
17767 similar way to the verify program. Tidy up the crl program so it
17768 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17769 less strict. It will now permit CRL extensions even if it is not
17770 a V2 CRL: this will allow it to tolerate some broken CRLs.
17774 * Initialize all non-automatic variables each time one of the openssl
17775 sub-programs is started (this is necessary as they may be started
17776 multiple times from the "OpenSSL>" prompt).
17778 *Lennart Bang, Bodo Moeller*
17780 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17781 removing all other RSA functionality (this is what NO_RSA does). This
17782 is so (for example) those in the US can disable those operations covered
17783 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17788 * Non-copying interface to BIO pairs.
17789 (still largely untested)
17793 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17794 ASCII string. This was handled independently in various places before.
17798 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17799 UTF8 strings a character at a time.
17803 * Use client_version from client hello to select the protocol
17804 (s23_srvr.c) and for RSA client key exchange verification
17805 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17809 * Add various utility functions to handle SPKACs, these were previously
17810 handled by poking round in the structure internals. Added new function
17811 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17812 print, verify and generate SPKACs. Based on an original idea from
17813 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17817 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17821 * Allow the config file extension section to be overwritten on the
17822 command line. Based on an original idea from Massimiliano Pala
17823 <madwolf@comune.modena.it>. The new option is called -extensions
17824 and can be applied to ca, req and x509. Also -reqexts to override
17825 the request extensions in req and -crlexts to override the crl extensions
17830 * Add new feature to the SPKAC handling in ca. Now you can include
17831 the same field multiple times by preceding it by "XXXX." for example:
17834 this is the same syntax as used in the req config file.
17838 * Allow certificate extensions to be added to certificate requests. These
17839 are specified in a 'req_extensions' option of the req section of the
17840 config file. They can be printed out with the -text option to req but
17841 are otherwise ignored at present.
17845 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17846 data read consists of only the final block it would not decrypted because
17847 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17848 A misplaced 'break' also meant the decrypted final block might not be
17849 copied until the next read.
17853 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17854 a few extra parameters to the DH structure: these will be useful if
17855 for example we want the value of 'q' or implement X9.42 DH.
17859 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17860 provides hooks that allow the default DSA functions or functions on a
17861 "per key" basis to be replaced. This allows hardware acceleration and
17862 hardware key storage to be handled without major modification to the
17863 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17864 associated functions.
17868 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17869 as "read only": it can't be written to and the buffer it points to will
17870 not be freed. Reading from a read only BIO is much more efficient than
17871 a normal memory BIO. This was added because there are several times when
17872 an area of memory needs to be read from a BIO. The previous method was
17873 to create a memory BIO and write the data to it, this results in two
17874 copies of the data and an O(n^2) reading algorithm. There is a new
17875 function BIO_new_mem_buf() which creates a read only memory BIO from
17876 an area of memory. Also modified the PKCS#7 routines to use read only
17881 * Bugfix: ssl23_get_client_hello did not work properly when called in
17882 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17883 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17884 but a retry condition occurred while trying to read the rest.
17888 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17889 NID_pkcs7_encrypted by default: this was wrong since this should almost
17890 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17891 the encrypted data type: this is a more sensible place to put it and it
17892 allows the PKCS#12 code to be tidied up that duplicated this
17897 * Changed obj_dat.pl script so it takes its input and output files on
17898 the command line. This should avoid shell escape redirection problems
17903 * Initial support for certificate extension requests, these are included
17904 in things like Xenroll certificate requests. Included functions to allow
17905 extensions to be obtained and added.
17909 * -crlf option to s_client and s_server for sending newlines as
17910 CRLF (as required by many protocols).
17914 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17916 * Install libRSAglue.a when OpenSSL is built with RSAref.
17918 *Ralf S. Engelschall*
17920 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17922 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17924 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17929 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17930 DH parameters/keys (q is lost during that conversion, but the resulting
17931 DH parameters contain its length).
17933 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17934 much faster than DH_generate_parameters (which creates parameters
17935 where `p = 2*q + 1`), and also the smaller q makes DH computations
17936 much more efficient (160-bit exponentiation instead of 1024-bit
17937 exponentiation); so this provides a convenient way to support DHE
17938 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17939 utter importance to use
17940 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17942 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17943 when such DH parameters are used, because otherwise small subgroup
17944 attacks may become possible!
17948 * Avoid memory leak in i2d_DHparams.
17952 * Allow the -k option to be used more than once in the enc program:
17953 this allows the same encrypted message to be read by multiple recipients.
17957 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17958 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17959 it will always use the numerical form of the OID, even if it has a short
17964 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17965 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17966 otherwise bn_mod_exp was called. In the case of hardware keys for example
17967 no private key components need be present and it might store extra data
17968 in the RSA structure, which cannot be accessed from bn_mod_exp.
17969 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17970 private key operations.
17974 * Added support for SPARC Linux.
17978 * pem_password_cb function type incompatibly changed from
17979 typedef int pem_password_cb(char *buf, int size, int rwflag);
17981 ....(char *buf, int size, int rwflag, void *userdata);
17982 so that applications can pass data to their callbacks:
17983 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17984 additional void * argument, which is just handed through whenever
17985 the password callback is called.
17987 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17989 New function SSL_CTX_set_default_passwd_cb_userdata.
17991 Compatibility note: As many C implementations push function arguments
17992 onto the stack in reverse order, the new library version is likely to
17993 interoperate with programs that have been compiled with the old
17994 pem_password_cb definition (PEM_whatever takes some data that
17995 happens to be on the stack as its last argument, and the callback
17996 just ignores this garbage); but there is no guarantee whatsoever that
17999 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18000 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18001 problems not only on Windows, but also on some Unix platforms.
18002 To avoid problematic command lines, these definitions are now in an
18003 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18004 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18008 * MIPS III/IV assembler module is reimplemented.
18012 * More DES library cleanups: remove references to srand/rand and
18013 delete an unused file.
18017 * Add support for the free Netwide assembler (NASM) under Win32,
18018 since not many people have MASM (ml) and it can be hard to obtain.
18019 This is currently experimental but it seems to work OK and pass all
18020 the tests. Check out INSTALL.W32 for info.
18024 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18025 without temporary keys kept an extra copy of the server key,
18026 and connections with temporary keys did not free everything in case
18031 * New function RSA_check_key and new openssl rsa option -check
18032 for verifying the consistency of RSA keys.
18034 *Ulf Moeller, Bodo Moeller*
18036 * Various changes to make Win32 compile work:
18037 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18038 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18039 comparison" warnings.
18040 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18044 * Add a debugging option to PKCS#5 v2 key generation function: when
18045 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18046 derived keys are printed to stderr.
18050 * Copy the flags in ASN1_STRING_dup().
18052 *Roman E. Pavlov <pre@mo.msk.ru>*
18054 * The x509 application mishandled signing requests containing DSA
18055 keys when the signing key was also DSA and the parameters didn't match.
18057 It was supposed to omit the parameters when they matched the signing key:
18058 the verifying software was then supposed to automatically use the CA's
18059 parameters if they were absent from the end user certificate.
18061 Omitting parameters is no longer recommended. The test was also
18062 the wrong way round! This was probably due to unusual behaviour in
18063 EVP_cmp_parameters() which returns 1 if the parameters match.
18064 This meant that parameters were omitted when they *didn't* match and
18065 the certificate was useless. Certificates signed with 'ca' didn't have
18068 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18070 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18071 The interface is as follows:
18072 Applications can use
18073 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18074 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18075 "off" is now the default.
18076 The library internally uses
18077 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18078 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18079 to disable memory-checking temporarily.
18081 Some inconsistent states that previously were possible (and were
18082 even the default) are now avoided.
18084 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18085 with each memory chunk allocated; this is occasionally more helpful
18086 than just having a counter.
18088 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18090 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18095 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18096 which largely parallels "options", but is for changing API behaviour,
18097 whereas "options" are about protocol behaviour.
18098 Initial "mode" flags are:
18100 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18101 a single record has been written.
18102 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18103 retries use the same buffer location.
18104 (But all of the contents must be
18109 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18112 * Fix problems with no-hmac etc.
18114 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18116 * New functions RSA_get_default_method(), RSA_set_method() and
18117 RSA_get_method(). These allows replacement of RSA_METHODs without having
18118 to mess around with the internals of an RSA structure.
18122 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18123 Also really enable memory leak checks in openssl.c and in some
18126 *Chad C. Mulligan, Bodo Moeller*
18128 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18129 up the length of negative integers. This has now been simplified to just
18130 store the length when it is first determined and use it later, rather
18131 than trying to keep track of where data is copied and updating it to
18133 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18135 * Add a new function PKCS7_signatureVerify. This allows the verification
18136 of a PKCS#7 signature but with the signing certificate passed to the
18137 function itself. This contrasts with PKCS7_dataVerify which assumes the
18138 certificate is present in the PKCS#7 structure. This isn't always the
18139 case: certificates can be omitted from a PKCS#7 structure and be
18140 distributed by "out of band" means (such as a certificate database).
18144 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18145 function prototypes in pem.h, also change util/mkdef.pl to add the
18146 necessary function names.
18150 * mk1mf.pl (used by Windows builds) did not properly read the
18151 options set by Configure in the top level Makefile, and Configure
18152 was not even able to write more than one option correctly.
18153 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18157 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18158 file to be loaded from a BIO or FILE pointer. The BIO version will
18159 for example allow memory BIOs to contain config info.
18163 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18164 Whoever hopes to achieve shared-library compatibility across versions
18165 must use this, not the compile-time macro.
18166 (Exercise 0.9.4: Which is the minimum library version required by
18168 Note: All this applies only to multi-threaded programs, others don't
18173 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18174 through a BIO pair triggered the default case, i.e.
18175 SSLerr(...,SSL_R_UNKNOWN_STATE).
18179 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18180 can use the SSL library even if none of the specific BIOs is
18185 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18186 for the encoded length.
18188 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18190 * Add initial documentation of the X509V3 functions.
18194 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18195 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18196 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18197 secure PKCS#8 private key format with a high iteration count.
18201 * Fix determination of Perl interpreter: A perl or perl5
18202 *directory* in $PATH was also accepted as the interpreter.
18204 *Ralf S. Engelschall*
18206 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18207 wrong with it but it was very old and did things like calling
18208 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18209 unusual formatting.
18213 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18214 to use the new extension code.
18218 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18219 with macros. This should make it easier to change their form, add extra
18220 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18225 * Add to configuration table a new entry that can specify an alternative
18226 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18227 according to Mark Crispin <MRC@Panda.COM>.
18231 * DES CBC did not update the IV. Weird.
18235 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18236 Changing the behaviour of the former might break existing programs --
18237 where IV updating is needed, des_ncbc_encrypt can be used.
18240 * When bntest is run from "make test" it drives bc to check its
18241 calculations, as well as internally checking them. If an internal check
18242 fails, it needs to cause bc to give a non-zero result or make test carries
18243 on without noticing the failure. Fixed.
18247 * DES library cleanups.
18251 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18252 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18253 ciphers. NOTE: although the key derivation function has been verified
18254 against some published test vectors it has not been extensively tested
18255 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18260 * Instead of "mkdir -p", which is not fully portable, use new
18261 Perl script "util/mkdir-p.pl".
18265 * Rewrite the way password based encryption (PBE) is handled. It used to
18266 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18267 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18268 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18269 the 'parameter' field of the AlgorithmIdentifier is passed to the
18270 underlying key generation function so it must do its own ASN1 parsing.
18271 This has also changed the EVP_PBE_CipherInit() function which now has a
18272 'parameter' argument instead of literal salt and iteration count values
18273 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18277 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18278 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18279 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18280 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18281 value was just used as a "magic string" and not used directly its
18282 value doesn't matter.
18286 * Introduce some semblance of const correctness to BN. Shame C doesn't
18291 * "linux-sparc64" configuration (ultrapenguin).
18293 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18294 "linux-sparc" configuration.
18296 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18298 * config now generates no-xxx options for missing ciphers.
18302 * Support the EBCDIC character set (work in progress).
18303 File ebcdic.c not yet included because it has a different license.
18305 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18307 * Support BS2000/OSD-POSIX.
18309 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18311 * Make callbacks for key generation use `void *` instead of `char *`.
18315 * Make S/MIME samples compile (not yet tested).
18319 * Additional typesafe stacks.
18323 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18327 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18329 * New configuration variant "sco5-gcc".
18331 * Updated some demos.
18333 *Sean O Riordain, Wade Scholine*
18335 * Add missing BIO_free at exit of pkcs12 application.
18339 * Fix memory leak in conf.c.
18343 * Updates for Win32 to assembler version of MD5.
18347 * Set #! path to perl in `apps/der_chop` to where we found it
18348 instead of using a fixed path.
18352 * SHA library changes for irix64-mips4-cc.
18356 * Improvements for VMS support.
18360 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18362 * Bignum library bug fix. IRIX 6 passes "make test" now!
18363 This also avoids the problems with SC4.2 and unpatched SC5.
18365 *Andy Polyakov <appro@fy.chalmers.se>*
18367 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18368 These are required because of the typesafe stack would otherwise break
18369 existing code. If old code used a structure member which used to be STACK
18370 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18371 sk_num or sk_value it would produce an error because the num, data members
18372 are not present in STACK_OF. Now it just produces a warning. sk_set
18373 replaces the old method of assigning a value to sk_value
18374 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18375 that does this will no longer work (and should use sk_set instead) but
18376 this could be regarded as a "questionable" behaviour anyway.
18380 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18381 correctly handle encrypted S/MIME data.
18385 * Change type of various DES function arguments from des_cblock
18386 (which means, in function argument declarations, pointer to char)
18387 to des_cblock * (meaning pointer to array with 8 char elements),
18388 which allows the compiler to do more typechecking; it was like
18389 that back in SSLeay, but with lots of ugly casts.
18391 Introduce new type const_des_cblock.
18395 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18396 problems: find RecipientInfo structure that matches recipient certificate
18397 and initialise the ASN1 structures properly based on passed cipher.
18401 * Belatedly make the BN tests actually check the results.
18405 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18406 to and from BNs: it was completely broken. New compilation option
18407 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18408 key elements as negative integers.
18412 * Reorganize and speed up MD5.
18414 *Andy Polyakov <appro@fy.chalmers.se>*
18418 *Richard Levitte <richard@levitte.org>*
18420 * New option -out to asn1parse to allow the parsed structure to be
18421 output to a file. This is most useful when combined with the -strparse
18422 option to examine the output of things like OCTET STRINGS.
18426 * Make SSL library a little more fool-proof by not requiring any longer
18427 that `SSL_set_{accept,connect}_state` be called before
18428 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18429 in many applications because usually everything *appeared* to work as
18430 intended anyway -- now it really works as intended).
18434 * Move openssl.cnf out of lib/.
18438 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18439 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18440 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18442 *Ralf S. Engelschall*
18444 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18445 handle PKCS#7 enveloped data properly.
18447 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18449 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18450 copying pointers. The cert_st handling is changed by this in
18451 various ways (and thus what used to be known as ctx->default_cert
18452 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18453 any longer when s->cert does not give us what we need).
18454 ssl_cert_instantiate becomes obsolete by this change.
18455 As soon as we've got the new code right (possibly it already is?),
18456 we have solved a couple of bugs of the earlier code where s->cert
18457 was used as if it could not have been shared with other SSL structures.
18459 Note that using the SSL API in certain dirty ways now will result
18460 in different behaviour than observed with earlier library versions:
18461 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18462 does not influence s as it used to.
18464 In order to clean up things more thoroughly, inside SSL_SESSION
18465 we don't use CERT any longer, but a new structure SESS_CERT
18466 that holds per-session data (if available); currently, this is
18467 the peer's certificate chain and, for clients, the server's certificate
18468 and temporary key. CERT holds only those values that can have
18469 meaningful defaults in an SSL_CTX.
18473 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18474 from the internal representation. Various PKCS#7 fixes: remove some
18475 evil casts and set the enc_dig_alg field properly based on the signing
18480 * Allow PKCS#12 password to be set from the command line or the
18481 environment. Let 'ca' get its config file name from the environment
18482 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18487 * Allow certificate policies extension to use an IA5STRING for the
18488 organization field. This is contrary to the PKIX definition but
18489 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18494 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18495 without disallowing inline assembler and the like for non-pedantic builds.
18499 * Support Borland C++ builder.
18501 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18507 * SHA-1 cleanups and performance enhancements.
18509 *Andy Polyakov <appro@fy.chalmers.se>*
18511 * Sparc v8plus assembler for the bignum library.
18513 *Andy Polyakov <appro@fy.chalmers.se>*
18515 * Accept any -xxx and +xxx compiler options in Configure.
18519 * Update HPUX configuration.
18523 * Add missing `sk_<type>_unshift()` function to safestack.h
18525 *Ralf S. Engelschall*
18527 * New function SSL_CTX_use_certificate_chain_file that sets the
18528 "extra_cert"s in addition to the certificate. (This makes sense
18529 only for "PEM" format files, as chains as a whole are not
18534 * Support verify_depth from the SSL API.
18535 x509_vfy.c had what can be considered an off-by-one-error:
18536 Its depth (which was not part of the external interface)
18537 was actually counting the number of certificates in a chain;
18538 now it really counts the depth.
18542 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18543 instead of X509err, which often resulted in confusing error
18544 messages since the error codes are not globally unique
18545 (e.g. an alleged error in ssl3_accept when a certificate
18546 didn't match the private key).
18548 * New function SSL_CTX_set_session_id_context that allows to set a default
18549 value (so that you don't need SSL_set_session_id_context for each
18550 connection using the SSL_CTX).
18554 * OAEP decoding bug fix.
18558 * Support INSTALL_PREFIX for package builders, as proposed by
18563 * New Configure options "threads" and "no-threads". For systems
18564 where the proper compiler options are known (currently Solaris
18565 and Linux), "threads" is the default.
18569 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18573 * Install various scripts to $(OPENSSLDIR)/misc, not to
18574 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18575 such as /usr/local/bin.
18579 * "make linux-shared" to build shared libraries.
18581 *Niels Poppe <niels@netbox.org>*
18583 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18587 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18588 extension adding in x509 utility.
18592 * Remove NOPROTO sections and error code comments.
18596 * Partial rewrite of the DEF file generator to now parse the ANSI
18601 * New Configure options --prefix=DIR and --openssldir=DIR.
18605 * Complete rewrite of the error code script(s). It is all now handled
18606 by one script at the top level which handles error code gathering,
18607 header rewriting and C source file generation. It should be much better
18608 than the old method: it now uses a modified version of Ulf's parser to
18609 read the ANSI prototypes in all header files (thus the old K&R definitions
18610 aren't needed for error creation any more) and do a better job of
18611 translating function codes into names. The old 'ASN1 error code embedded
18612 in a comment' is no longer necessary and it doesn't use .err files which
18613 have now been deleted. Also the error code call doesn't have to appear all
18614 on one line (which resulted in some large lines...).
18618 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18622 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18623 0 (which usually indicates a closed connection), but continue reading.
18627 * Fix some race conditions.
18631 * Add support for CRL distribution points extension. Add Certificate
18632 Policies and CRL distribution points documentation.
18636 * Move the autogenerated header file parts to crypto/opensslconf.h.
18640 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18641 8 of keying material. Merlin has also confirmed interop with this fix
18642 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18644 *Merlin Hughes <merlin@baltimore.ie>*
18646 * Fix lots of warnings.
18648 *Richard Levitte <levitte@stacken.kth.se>*
18650 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18651 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18653 *Richard Levitte <levitte@stacken.kth.se>*
18655 * Fix problems with sizeof(long) == 8.
18657 *Andy Polyakov <appro@fy.chalmers.se>*
18659 * Change functions to ANSI C.
18663 * Fix typos in error codes.
18665 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18667 * Remove defunct assembler files from Configure.
18671 * SPARC v8 assembler BIGNUM implementation.
18673 *Andy Polyakov <appro@fy.chalmers.se>*
18675 * Support for Certificate Policies extension: both print and set.
18676 Various additions to support the r2i method this uses.
18680 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18681 return a const string when you are expecting an allocated buffer.
18685 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18686 types DirectoryString and DisplayText.
18690 * Add code to allow r2i extensions to access the configuration database,
18691 add an LHASH database driver and add several ctx helper functions.
18695 * Fix an evil bug in bn_expand2() which caused various BN functions to
18696 fail when they extended the size of a BIGNUM.
18700 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18701 support typesafe stack.
18705 * Fix typo in SSL_[gs]et_options().
18707 *Nils Frostberg <nils@medcom.se>*
18709 * Delete various functions and files that belonged to the (now obsolete)
18710 old X509V3 handling code.
18714 * New Configure option "rsaref".
18718 * Don't auto-generate pem.h.
18722 * Introduce type-safe ASN.1 SETs.
18726 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18728 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18730 * Introduce type-safe STACKs. This will almost certainly break lots of code
18731 that links with OpenSSL (well at least cause lots of warnings), but fear
18732 not: the conversion is trivial, and it eliminates loads of evil casts. A
18733 few STACKed things have been converted already. Feel free to convert more.
18734 In the fullness of time, I'll do away with the STACK type altogether.
18738 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18739 specified in `<certfile>` by updating the entry in the index.txt file.
18740 This way one no longer has to edit the index.txt file manually for
18741 revoking a certificate. The -revoke option does the gory details now.
18743 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18745 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18746 `-text` option at all and this way the `-noout -text` combination was
18747 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18749 *Ralf S. Engelschall*
18751 * Make sure a corresponding plain text error message exists for the
18752 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18753 verify callback function determined that a certificate was revoked.
18755 *Ralf S. Engelschall*
18757 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18758 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18759 all available ciphers including rc5, which was forgotten until now.
18760 In order to let the testing shell script know which algorithms
18761 are available, a new (up to now undocumented) command
18762 `openssl list-cipher-commands` is used.
18766 * Bugfix: s_client occasionally would sleep in select() when
18767 it should have checked SSL_pending() first.
18771 * New functions DSA_do_sign and DSA_do_verify to provide access to
18772 the raw DSA values prior to ASN.1 encoding.
18776 * Tweaks to Configure
18778 *Niels Poppe <niels@netbox.org>*
18780 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18785 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18789 * New config option to avoid instructions that are illegal on the 80386.
18790 The default code is faster, but requires at least a 486.
18794 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18795 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18796 same as SSL2_VERSION anyway.
18800 * New "-showcerts" option for s_client.
18804 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18805 application. Various cleanups and fixes.
18809 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18810 modify error routines to work internally. Add error codes and PBE init
18811 to library startup routines.
18815 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18816 packing functions to asn1 and evp. Changed function names and error
18817 codes along the way.
18821 * PKCS12 integration: and so it begins... First of several patches to
18822 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18823 objects to objects.h
18827 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18828 and display support for Thawte strong extranet extension.
18832 * Add LinuxPPC support.
18834 *Jeff Dubrule <igor@pobox.org>*
18836 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18837 bn_div_words in alpha.s.
18839 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18841 * Make sure the RSA OAEP test is skipped under -DRSAref because
18842 OAEP isn't supported when OpenSSL is built with RSAref.
18844 *Ulf Moeller <ulf@fitug.de>*
18846 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18847 so they no longer are missing under -DNOPROTO.
18849 *Soren S. Jorvang <soren@t.dk>*
18851 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18853 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18854 doesn't work when the session is reused. Coming soon!
18858 * Fix a security hole, that allows sessions to be reused in the wrong
18859 context thus bypassing client cert protection! All software that uses
18860 client certs and session caches in multiple contexts NEEDS PATCHING to
18861 allow session reuse! A fuller solution is in the works.
18863 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18865 * Some more source tree cleanups (removed obsolete files
18866 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18867 permission on "config" script to be executable) and a fix for the INSTALL
18870 *Ulf Moeller <ulf@fitug.de>*
18872 * Remove some legacy and erroneous uses of malloc, free instead of
18875 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18877 * Make rsa_oaep_test return non-zero on error.
18879 *Ulf Moeller <ulf@fitug.de>*
18881 * Add support for native Solaris shared libraries. Configure
18882 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18883 if someone would make that last step automatic.
18885 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18887 * ctx_size was not built with the right compiler during "make links". Fixed.
18891 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18892 except NULL ciphers". This means the default cipher list will no longer
18893 enable NULL ciphers. They need to be specifically enabled e.g. with
18894 the string "DEFAULT:eNULL".
18898 * Fix to RSA private encryption routines: if p < q then it would
18899 occasionally produce an invalid result. This will only happen with
18900 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18904 * Be less restrictive and allow also `perl util/perlpath.pl
18905 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18906 because this way one can also use an interpreter named `perl5` (which is
18907 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18908 installed as `perl`).
18910 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18912 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18914 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18916 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18917 advapi32.lib to Win32 build and change the pem test comparison
18918 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18919 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18920 and crypto/des/ede_cbcm_enc.c.
18924 * DES quad checksum was broken on big-endian architectures. Fixed.
18928 * Comment out two functions in bio.h that aren't implemented. Fix up the
18929 Win32 test batch file so it (might) work again. The Win32 test batch file
18930 is horrible: I feel ill....
18934 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18935 in e_os.h. Audit of header files to check ANSI and non ANSI
18936 sections: 10 functions were absent from non ANSI section and not exported
18937 from Windows DLLs. Fixed up libeay.num for new functions.
18941 * Make `openssl version` output lines consistent.
18943 *Ralf S. Engelschall*
18945 * Fix Win32 symbol export lists for BIO functions: Added
18946 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18947 to ms/libeay{16,32}.def.
18949 *Ralf S. Engelschall*
18951 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18952 fine under Unix and passes some trivial tests I've now added. But the
18953 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18954 added to make sure no one expects that this stuff really works in the
18955 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18956 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18959 *Ralf S. Engelschall*
18961 * Fix the generation of two part addresses in perl.
18963 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18965 * Add config entry for Linux on MIPS.
18967 *John Tobey <jtobey@channel1.com>*
18969 * Make links whenever Configure is run, unless we are on Windoze.
18973 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18974 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18979 * Add a useful kludge to allow package maintainers to specify compiler and
18980 other platforms details on the command line without having to patch the
18981 Configure script every time: One now can use
18982 `perl Configure <id>:<details>`,
18983 i.e. platform ids are allowed to have details appended
18984 to them (separated by colons). This is treated as there would be a static
18985 pre-configured entry in Configure's %table under key `<id>` with value
18986 `<details>` and `perl Configure <id>` is called. So, when you want to
18987 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18988 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18989 now, which overrides the FreeBSD-elf entry on-the-fly.
18991 *Ralf S. Engelschall*
18993 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18997 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18998 on the `perl Configure ...` command line. This way one can compile
18999 OpenSSL libraries with Position Independent Code (PIC) which is needed
19000 for linking it into DSOs.
19002 *Ralf S. Engelschall*
19004 * Remarkably, export ciphers were totally broken and no-one had noticed!
19009 * Cleaned up the LICENSE document: The official contact for any license
19010 questions now is the OpenSSL core team under openssl-core@openssl.org.
19011 And add a paragraph about the dual-license situation to make sure people
19012 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19013 to the OpenSSL toolkit.
19015 *Ralf S. Engelschall*
19017 * General source tree makefile cleanups: Made `making xxx in yyy...`
19018 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19019 Additionally cleaned up the `make links` target: Remove unnecessary
19020 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19021 to speed processing and no longer clutter the display with confusing
19022 stuff. Instead only the actually done links are displayed.
19024 *Ralf S. Engelschall*
19026 * Permit null encryption ciphersuites, used for authentication only. It used
19027 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19028 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19033 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19034 signed attributes when verifying signatures (this would break them),
19035 the detached data encoding was wrong and public keys obtained using
19036 X509_get_pubkey() weren't freed.
19040 * Add text documentation for the BUFFER functions. Also added a work around
19041 to a Win95 console bug. This was triggered by the password read stuff: the
19042 last character typed gets carried over to the next fread(). If you were
19043 generating a new cert request using 'req' for example then the last
19044 character of the passphrase would be CR which would then enter the first
19049 * Added the new 'Includes OpenSSL Cryptography Software' button as
19050 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19051 button and can be used by applications based on OpenSSL to show the
19052 relationship to the OpenSSL project.
19054 *Ralf S. Engelschall*
19056 * Remove confusing variables in function signatures in files
19057 ssl/ssl_lib.c and ssl/ssl.h.
19059 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19061 * Don't install bss_file.c under PREFIX/include/
19063 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19065 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19066 functions that return function pointers and has support for NT specific
19067 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19068 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19069 unsigned to signed types: this was killing the Win32 compile.
19073 * Add new certificate file to stack functions,
19074 SSL_add_dir_cert_subjects_to_stack() and
19075 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19076 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19077 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19078 This means that Apache-SSL and similar packages don't have to mess around
19079 to add as many CAs as they want to the preferred list.
19083 * Experiment with doxygen documentation. Currently only partially applied to
19085 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19086 openssl.doxy as the configuration file.
19090 * Get rid of remaining C++-style comments which strict C compilers hate.
19092 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19094 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19095 compiled in by default: it has problems with large keys.
19099 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19100 DH private keys and/or callback functions which directly correspond to
19101 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19102 is needed for applications which have to configure certificates on a
19103 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19105 For the RSA certificate situation is makes no difference, but
19106 for the DSA certificate situation this fixes the "no shared cipher"
19107 problem where the OpenSSL cipher selection procedure failed because the
19108 temporary keys were not overtaken from the context and the API provided
19109 no way to reconfigure them.
19110 The new functions now let applications reconfigure the stuff and they
19111 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19112 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19113 non-public-API function ssl_cert_instantiate() is used as a helper
19114 function and also to reduce code redundancy inside ssl_rsa.c.
19116 *Ralf S. Engelschall*
19118 * Move s_server -dcert and -dkey options out of the undocumented feature
19119 area because they are useful for the DSA situation and should be
19120 recognized by the users.
19122 *Ralf S. Engelschall*
19124 * Fix the cipher decision scheme for export ciphers: the export bits are
19125 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19126 SSL_EXP_MASK. So, the original variable has to be used instead of the
19127 already masked variable.
19129 *Richard Levitte <levitte@stacken.kth.se>*
19131 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19133 *Richard Levitte <levitte@stacken.kth.se>*
19135 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19136 from `int` to `unsigned int` because it is a length and initialized by
19137 EVP_DigestFinal() which expects an `unsigned int *`.
19139 *Richard Levitte <levitte@stacken.kth.se>*
19141 * Don't hard-code path to Perl interpreter on shebang line of Configure
19142 script. Instead use the usual Shell->Perl transition trick.
19144 *Ralf S. Engelschall*
19146 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19147 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19148 -noout -modulus` as it's already the case for `openssl rsa -noout
19149 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19150 currently the public key is printed (a decision which was already done by
19151 `openssl dsa -modulus` in the past) which serves a similar purpose.
19152 Additionally the NO_RSA no longer completely removes the whole -modulus
19153 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19156 *Ralf S. Engelschall*
19158 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19159 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19161 *Arne Ansper <arne@ats.cyber.ee>*
19163 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19164 to be added. Now both 'req' and 'ca' can use new objects defined in the
19169 * Add cool BIO that does syslog (or event log on NT).
19171 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19173 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19174 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19175 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19176 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19180 * Add preliminary config info for new extension code.
19184 * Make RSA_NO_PADDING really use no padding.
19186 *Ulf Moeller <ulf@fitug.de>*
19188 * Generate errors when private/public key check is done.
19192 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19193 for some CRL extensions and new objects added.
19197 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19198 key usage extension and fuller support for authority key id.
19202 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19203 padding method for RSA, which is recommended for new applications in PKCS
19204 #1 v2.0 (RFC 2437, October 1998).
19205 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19206 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19207 against Bleichbacher's attack on RSA.
19208 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19211 * Updates to the new SSL compression code
19213 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19215 * Fix so that the version number in the master secret, when passed
19216 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19217 (because the server will not accept higher), that the version number
19218 is 0x03,0x01, not 0x03,0x00
19220 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19222 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19223 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19224 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19228 * Support for RAW extensions where an arbitrary extension can be
19229 created by including its DER encoding. See `apps/openssl.cnf` for
19234 * Make sure latest Perl versions don't interpret some generated C array
19235 code as Perl array code in the crypto/err/err_genc.pl script.
19237 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19239 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19240 not many people have the assembler. Various Win32 compilation fixes and
19241 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19242 build instructions.
19246 * Modify configure script 'Configure' to automatically create crypto/date.h
19247 file under Win32 and also build pem.h from pem.org. New script
19248 util/mkfiles.pl to create the MINFO file on environments that can't do a
19249 'make files': perl util/mkfiles.pl >MINFO should work.
19253 * Major rework of DES function declarations, in the pursuit of correctness
19254 and purity. As a result, many evil casts evaporated, and some weirdness,
19255 too. You may find this causes warnings in your code. Zapping your evil
19256 casts will probably fix them. Mostly.
19260 * Fix for a typo in asn1.h. Bug fix to object creation script
19261 obj_dat.pl. It considered a zero in an object definition to mean
19262 "end of object": none of the objects in objects.h have any zeros
19263 so it wasn't spotted.
19265 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19267 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19268 Masking (CBCM). In the absence of test vectors, the best I have been able
19269 to do is check that the decrypt undoes the encrypt, so far. Send me test
19270 vectors if you have them.
19274 * Correct calculation of key length for export ciphers (too much space was
19275 allocated for null ciphers). This has not been tested!
19279 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19280 message is now correct (it understands "crypto" and "ssl" on its
19281 command line). There is also now an "update" option. This will update
19282 the util/ssleay.num and util/libeay.num files with any new functions.
19284 perl util/mkdef.pl crypto ssl update
19285 it will update them.
19289 * Overhauled the Perl interface:
19290 - ported BN stuff to OpenSSL's different BN library
19291 - made the perl/ source tree CVS-aware
19292 - renamed the package from SSLeay to OpenSSL (the files still contain
19293 their history because I've copied them in the repository)
19294 - removed obsolete files (the test scripts will be replaced
19295 by better Test::Harness variants in the future)
19297 *Ralf S. Engelschall*
19299 * First cut for a very conservative source tree cleanup:
19300 1. merge various obsolete readme texts into doc/ssleay.txt
19301 where we collect the old documents and readme texts.
19302 2. remove the first part of files where I'm already sure that we no
19303 longer need them because of three reasons: either they are just temporary
19304 files which were left by Eric or they are preserved original files where
19305 I've verified that the diff is also available in the CVS via "cvs diff
19306 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19307 the crypto/md/ stuff).
19309 *Ralf S. Engelschall*
19311 * More extension code. Incomplete support for subject and issuer alt
19312 name, issuer and authority key id. Change the i2v function parameters
19313 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19314 what that's for :-) Fix to ASN1 macro which messed up
19315 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19319 * Preliminary support for ENUMERATED type. This is largely copied from the
19324 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19326 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19328 * Make sure `make rehash` target really finds the `openssl` program.
19330 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19332 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19333 like to hear about it if this slows down other processors.
19337 * Add CygWin32 platform information to Configure script.
19339 *Alan Batie <batie@aahz.jf.intel.com>*
19341 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19343 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19345 * New program nseq to manipulate netscape certificate sequences
19349 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19354 * Fixes to BN code. Previously the default was to define BN_RECURSION
19355 but the BN code had some problems that would cause failures when
19356 doing certificate verification and some other functions.
19358 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19360 * Add ASN1 and PEM code to support netscape certificate sequences.
19364 * Add ASN1 and PEM code to support netscape certificate sequences.
19368 * Add several PKIX and private extended key usage OIDs.
19372 * Modify the 'ca' program to handle the new extension code. Modify
19373 openssl.cnf for new extension format, add comments.
19377 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19378 and add a sample to openssl.cnf so req -x509 now adds appropriate
19383 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19384 error code, add initial support to X509_print() and x509 application.
19388 * Takes a deep breath and start adding X509 V3 extension support code. Add
19389 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19390 stuff is currently isolated and isn't even compiled yet.
19394 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19395 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19396 Removed the versions check from X509 routines when loading extensions:
19397 this allows certain broken certificates that don't set the version
19398 properly to be processed.
19402 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19403 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19404 can still be regenerated with "make depend".
19408 * Spelling mistake in C version of CAST-128.
19410 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19412 * Changes to the error generation code. The perl script err-code.pl
19413 now reads in the old error codes and retains the old numbers, only
19414 adding new ones if necessary. It also only changes the .err files if new
19415 codes are added. The makefiles have been modified to only insert errors
19416 when needed (to avoid needlessly modifying header files). This is done
19417 by only inserting errors if the .err file is newer than the auto generated
19418 C file. To rebuild all the error codes from scratch (the old behaviour)
19419 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19420 or delete all the .err files.
19424 * CAST-128 was incorrectly implemented for short keys. The C version has
19425 been fixed, but is untested. The assembler versions are also fixed, but
19426 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19427 to regenerate it if needed.
19428 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19429 Hagino <itojun@kame.net>*
19431 * File was opened incorrectly in randfile.c.
19433 *Ulf Möller <ulf@fitug.de>*
19435 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19436 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19437 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19438 al: it's just almost always a UTCTime. Note this patch adds new error
19439 codes so do a "make errors" if there are problems.
19443 * Correct Linux 1 recognition in config.
19445 *Ulf Möller <ulf@fitug.de>*
19447 * Remove pointless MD5 hash when using DSA keys in ca.
19449 *Anonymous <nobody@replay.com>*
19451 * Generate an error if given an empty string as a cert directory. Also
19452 generate an error if handed NULL (previously returned 0 to indicate an
19453 error, but didn't set one).
19455 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19457 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19461 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19462 parameters. This was causing a warning which killed off the Win32 compile.
19466 * Remove C++ style comments from crypto/bn/bn_local.h.
19468 *Neil Costigan <neil.costigan@celocom.com>*
19470 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19471 based on a text string, looking up short and long names and finally
19472 "dot" format. The "dot" format stuff didn't work. Added new function
19473 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19474 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19475 OID is not part of the table.
19479 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19480 X509_LOOKUP_by_alias().
19484 * Sort openssl functions by name.
19488 * Get the `gendsa` command working and add it to the `list` command. Remove
19489 encryption from sample DSA keys (in case anyone is interested the password
19494 * Make *all* `*_free` functions accept a NULL pointer.
19496 *Frans Heymans <fheymans@isaserver.be>*
19498 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19501 *Anonymous <nobody@replay.com>*
19503 * s_server should send the CAfile as acceptable CAs, not its own cert.
19505 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19507 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19509 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19511 * Temp key "for export" tests were wrong in s3_srvr.c.
19513 *Anonymous <nobody@replay.com>*
19515 * Add prototype for temp key callback functions
19516 SSL_CTX_set_tmp_{rsa,dh}_callback().
19520 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19521 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19525 * X509_name_add_entry() freed the wrong thing after an error.
19527 *Arne Ansper <arne@ats.cyber.ee>*
19529 * rsa_eay.c would attempt to free a NULL context.
19531 *Arne Ansper <arne@ats.cyber.ee>*
19533 * BIO_s_socket() had a broken should_retry() on Windoze.
19535 *Arne Ansper <arne@ats.cyber.ee>*
19537 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19539 *Arne Ansper <arne@ats.cyber.ee>*
19541 * Make sure the already existing X509_STORE->depth variable is initialized
19542 in X509_STORE_new(), but document the fact that this variable is still
19543 unused in the certificate verification process.
19545 *Ralf S. Engelschall*
19547 * Fix the various library and `apps/` files to free up pkeys obtained from
19548 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19552 * Fix reference counting in X509_PUBKEY_get(). This makes
19553 demos/maurice/example2.c work, amongst others, probably.
19555 *Steve Henson and Ben Laurie*
19557 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19558 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19559 are no longer created. This way we have a single and consistent command
19560 line interface `openssl <command>`, similar to `cvs <command>`.
19562 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19564 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19565 BIT STRING wrapper always have zero unused bits.
19569 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19573 * Make the top-level INSTALL documentation easier to understand.
19577 * Makefiles updated to exit if an error occurs in a sub-directory
19578 make (including if user presses ^C) [Paul Sutton]
19580 * Make Montgomery context stuff explicit in RSA data structure.
19584 * Fix build order of pem and err to allow for generated pem.h.
19588 * Fix renumbering bug in X509_NAME_delete_entry().
19592 * Enhanced the err-ins.pl script so it makes the error library number
19593 global and can add a library name. This is needed for external ASN1 and
19594 other error libraries.
19598 * Fixed sk_insert which never worked properly.
19602 * Fix ASN1 macros so they can handle indefinite length constructed
19603 EXPLICIT tags. Some non standard certificates use these: they can now
19608 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19609 into a single doc/ssleay.txt bundle. This way the information is still
19610 preserved but no longer messes up this directory. Now it's new room for
19611 the new set of documentation files.
19613 *Ralf S. Engelschall*
19615 * SETs were incorrectly DER encoded. This was a major pain, because they
19616 shared code with SEQUENCEs, which aren't coded the same. This means that
19617 almost everything to do with SETs or SEQUENCEs has either changed name or
19618 number of arguments.
19620 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19622 * Fix test data to work with the above.
19626 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19627 was already fixed by Eric for 0.9.1 it seems.
19629 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19631 * Autodetect FreeBSD3.
19635 * Fix various bugs in Configure. This affects the following platforms:
19639 unixware-2.0-pentium
19644 * Eliminate generated files from CVS. Reorder tests to regenerate files
19645 before they are needed.
19649 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19653 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19655 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19656 changed SSLeay to OpenSSL in version strings.
19658 *Ralf S. Engelschall*
19660 * Some fixups to the top-level documents.
19664 * Fixed the nasty bug where rsaref.h was not found under compile-time
19665 because the symlink to include/ was missing.
19667 *Ralf S. Engelschall*
19669 * Incorporated the popular no-RSA/DSA-only patches
19670 which allow to compile an RSA-free SSLeay.
19672 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19674 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19675 when "ssleay" is still not found.
19677 *Ralf S. Engelschall*
19679 * Added more platforms to Configure: Cray T3E, HPUX 11,
19681 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19683 * Updated the README file.
19685 *Ralf S. Engelschall*
19687 * Added various .cvsignore files in the CVS repository subdirs
19688 to make a "cvs update" really silent.
19690 *Ralf S. Engelschall*
19692 * Recompiled the error-definition header files and added
19693 missing symbols to the Win32 linker tables.
19695 *Ralf S. Engelschall*
19697 * Cleaned up the top-level documents;
19698 o new files: CHANGES and LICENSE
19699 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19700 o merged COPYRIGHT into LICENSE
19701 o removed obsolete TODO file
19702 o renamed MICROSOFT to INSTALL.W32
19704 *Ralf S. Engelschall*
19706 * Removed dummy files from the 0.9.1b source tree:
19707 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19708 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19709 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19710 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19711 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19713 *Ralf S. Engelschall*
19715 * Added various platform portability fixes.
19719 * The Genesis of the OpenSSL rpject:
19720 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19721 Young and Tim J. Hudson created while they were working for C2Net until
19724 *The OpenSSL Project*
19726 ### Changes between 0.9.0b and 0.9.1b [not released]
19728 * Updated a few CA certificates under certs/
19732 * Changed some BIGNUM api stuff.
19736 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19737 DGUX x86, Linux Alpha, etc.
19741 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19742 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19747 * Add -strparse option to asn1pars program which parses nested
19750 *Dr Stephen Henson <shenson@bigfoot.com>*
19752 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19756 * DSA fix for "ca" program.
19760 * Added "-genkey" option to "dsaparam" program.
19764 * Added RIPE MD160 (rmd160) message digest.
19768 * Added -a (all) option to "ssleay version" command.
19772 * Added PLATFORM define which is the id given to Configure.
19776 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19780 * Extended the ASN.1 parser routines.
19784 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19788 * Added a BN_CTX to the BN library.
19792 * Fixed the weak key values in DES library
19796 * Changed API in EVP library for cipher aliases.
19800 * Added support for RC2/64bit cipher.
19804 * Converted the lhash library to the crypto/mem.c functions.
19808 * Added more recognized ASN.1 object ids.
19812 * Added more RSA padding checks for SSL/TLS.
19816 * Added BIO proxy/filter functionality.
19820 * Added extra_certs to SSL_CTX which can be used
19821 send extra CA certificates to the client in the CA cert chain sending
19822 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19826 * Now Fortezza is denied in the authentication phase because
19827 this is key exchange mechanism is not supported by SSLeay at all.
19831 * Additional PKCS1 checks.
19835 * Support the string "TLSv1" for all TLS v1 ciphers.
19839 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19840 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19844 * Fixed a few memory leaks.
19848 * Fixed various code and comment typos.
19852 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19853 bytes sent in the client random.
19855 *Edward Bishop <ebishop@spyglass.com>*
19859 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
19860 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
19861 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
19862 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
19863 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
19864 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
19865 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19866 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19867 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19868 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19869 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19870 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19871 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19872 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19873 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19874 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19875 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19876 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19877 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19878 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19879 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19880 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19881 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19882 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19883 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19884 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19885 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19886 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19887 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19888 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19889 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19890 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19891 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19892 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19893 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19894 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19895 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19896 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19897 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19898 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19899 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19900 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19901 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19902 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19903 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19904 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19905 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19906 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19907 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19908 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19909 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19910 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19911 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19912 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19913 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19914 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19915 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19916 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19917 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19918 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19919 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19920 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19921 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19922 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19923 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19924 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19925 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19926 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19927 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19928 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19929 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19930 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19931 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19932 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19933 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19934 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19935 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19936 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19937 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19938 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19939 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19940 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19941 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19942 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19943 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19944 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19945 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19946 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19947 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19948 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19949 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19950 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19951 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19952 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19953 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19954 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19955 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19956 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19957 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19958 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19959 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19960 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19961 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19962 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19963 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19964 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19965 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19966 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19967 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19968 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19969 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19970 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19971 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19972 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19973 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19974 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19975 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19976 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19977 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19978 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19979 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19980 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19981 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19982 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19983 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19984 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19985 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19986 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19987 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19988 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19989 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19990 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19991 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19992 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19993 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19994 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19995 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19996 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19997 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19998 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19999 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20000 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20001 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20002 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20003 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20004 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20005 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20006 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20007 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20008 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20009 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20010 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20011 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20012 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20013 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20014 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20015 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20016 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20017 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20018 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20019 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20020 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20021 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20022 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20023 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20024 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20025 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20026 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20027 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20028 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20029 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20030 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20031 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20032 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20033 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20034 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20035 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20036 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20037 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20038 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20039 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20040 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20041 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655